June 24, 2026
Time	Location	Event
12:25 to 12:40	Room 1	P0: Opening, Room 1
12:40 to 14:20	see below	C1: Contributed session 1, Rooms 1, 2, 3, 4
14:20 to 14:50	PJ floor and rooms PJ15/PJ16	Coffee break, PJ floor and rooms PJ15/PJ16
14:50 to 16:30	see below	C2: Contributed session 2, Rooms 1, 2, 3, 4
16:30 to 17:45	Room 1	P1: Keynote Michela Giorcelli: “Business School Education and Career Outcomes”, Room 1
19:00 to 22:00	Boat trip	Conference social dinner, Boat trip (Le Capitaine FRACASSE) , Boat trip
June 25, 2026
Time	Location	Event
09:00 to 10:40	see below	C3: contributed session 3, Rooms 1, 2, 3, 4
10:40 to 11:10	PJ floor and rooms PJ15/PJ16	Coffee break, PJ floor and rooms PJ15/PJ16
11:10 to 12:25	see below	C4: Contributed session 4, Rooms 1, 2, 3
12:25 to 13:30	PJ floor and rooms PJ15/PJ16	Lunch, PJ floor and rooms PJ15/PJ16
13:30 to 14:45	Room 1	P2: Keynote Pedro Sant'Anna “Efficient Difference-in-Differences and Event Study Estimators”, joint work with Xiaohong Chen and Haitian Xie, Room 1
14:45 to 16:00	see below	C5: Contributed session 5, Rooms 1, 2, 3
16:00 to 16:30	PJ floor and rooms PJ15/PJ16	Coffee break, PJ floor and rooms PJ15/PJ16
16:30 to 17:45	see below	C6: Contributed session 6, Rooms 1, 2, 3
18:00 to 19:00		Cocktail reception at the Les Gentlemens
June 26, 2026
Time	Location	Event
09:00 to 10:40	see below	C7: Contributed session 7, Rooms 1, 2, 3, 4
10:40 to 11:10	PJ floor and rooms PJ15/PJ16	Coffee break, PJ floor and rooms PJ15/PJ16
11:10 to 12:25	see below	C8: Contributed session 8, Rooms 1, 2, 3
12:25 to 13:20	PJ floor and rooms PJ15/PJ16	Lunch, PJ floor and rooms PJ15/PJ16
13:20 to 15:00	Room 1	P3: Roundtable “Causal analysis for business decisions: how can AI help?” & Closing, Room 1

Summary of All Sessions
Click here for an index of all participants

#	Date/Time	Type	Title/Location	Papers
1	June 24, 2026 12:40-14:20	contributed	Skills & training     Location: Room 1	4
2	June 24, 2026 12:40-14:20	contributed	Health     Location: Room 2	4
3	June 24, 2026 12:40-14:20	contributed	Education 1     Location: Room 3	4
4	June 24, 2026 12:40-14:20	contributed	Regional growth & transport     Location: Room 4	3
5	June 24, 2026 14:50-16:30	contributed	Labour markets     Location: Room 4	4
6	June 24, 2026 14:50-16:30	contributed	Health & aging     Location: Room 2	4
7	June 24, 2026 14:50-16:30	contributed	Higher education 1     Location: Room 3	4
8	June 24, 2026 14:50-16:30	contributed	Labour 1     Location: Room 1	4
9	June 25, 2026 9:00-10:40	contributed	Pay transparency & female labour supply     Location: Room 1	4
10	June 25, 2026 9:00-10:40	contributed	Education 2     Location: Room 2	4
11	June 25, 2026 9:00-10:40	contributed	Causal methods & learning     Location: Room 3	4
12	June 25, 2026 9:00-10:40	contributed	Place-based policies     Location: Room 4	4
13	June 25, 2026 11:10-12:25	contributed	Development     Location: Room 1	3
14	June 25, 2026 11:10-12:25	contributed	Firms     Location: Room 2	3
15	June 25, 2026 11:10-12:25	contributed	Causal methods & Applications     Location: Room 3	3
16	June 25, 2026 11:10-12:25	contributed	Shocks & labour markets     Location: Room 4	3
17	June 25, 2026 14:45-16:00	contributed	Migration     Location: Room 1	3
18	June 25, 2026 14:45-16:00	contributed	Firms' finances     Location: Room 2	3
19	June 25, 2026 14:45-16:00	contributed	Application of causal methods     Location: Room 3	3
20	June 25, 2026 16:30-17:45	contributed	Intergenerational     Location: Room 1	2
21	June 25, 2026 16:30-17:45	contributed	Environment 1     Location: Room 3	3
22	June 25, 2026 16:30-17:45	contributed	Higher Education 2     Location: Room 2	3
23	June 26, 2026 9:00-10:40	contributed	Wages     Location: Room 1	3
24	June 26, 2026 9:00-10:40	contributed	Labour 2     Location: Room 2	4
25	June 26, 2026 9:00-10:40	contributed	Shocks & mobility     Location: Room 3	4
26	June 26, 2026 9:00-10:40	contributed	Public good     Location: Room 4	4
27	June 26, 2026 11:10-12:25	contributed	Labour & shocks     Location: Room 1	3
28	June 26, 2026 11:10-12:25	contributed	ALMP     Location: Room 2	3
29	June 26, 2026 11:10-12:25	contributed	Environment 2     Location: Room 3	3

 

29 sessions, 100 papers, and 0 presentations with no associated papers

COunterfactual Methods for Policy Impact Evaluation 2026 (COMPIE 2026)
Detailed List of Sessions

Session 1: Skills & training June 24, 2026 12:40 to 14:20 Location: Room 1
Session Chair: Carlos Victoria-Lanzon, Universidad Complutense de Madrid
Session type: contributed
The effects of online vs. in-person instruction in subsidized further training on labour market outcomes Abstract Online instruction in subsidized further training has rapidly expanded in the last few years, yet little is known about the relative effectiveness of online vs. face-to-face teaching in improving unemployed participants’ labour market outcomes. Combining survey and administrative data and applying an event-study approach with inverse probability weighting, we analyse the employment and earnings effects of an exogenous and unanticipated switch from in-person to online instruction. We find a moderate, but somewhat imprecisely estimated positive effect of online vs. in-person teaching on participants’ probability to be employed in the three years following training begin. This is primarily driven by an increase in fulltime employment which also translates into significantly higher quarterly earnings starting six months after training begin. Heterogeneity analyses show that particularly men and higher-educated individuals benefit from online instruction, though the effects for men fade out over time. Overall, our results suggest that online courses can be an effective means to expand access to subsidized further training.
Presented by: Elisabeth Artmann, Institute for Employment Research
Wage Subsidies, Human Capital, and Formalization: Lessons from Bolivia’s Program to Support Employment II Abstract Can a temporary wage subsidy produce effects on formal employment that outlast the subsidy itself? We study Bolivia's Program to Support Employment~II, which combined three-month wage subsidies, job matching, and on-the-job training for low-skilled workers between 2018 and 2023. In a labor market where over 80\% of workers are informal and regulatory hiring costs are among the highest in Latin America, information frictions prevent productive workers from accessing formal jobs. Using difference-in-differences methods, we find persistent increases in formal employment and income that last well beyond the subsidy period, with larger effects for women than for men. Guided by a model of hiring under incomplete information, we distinguish two main channels through which the program may generate lasting effects---human capital accumulation and signaling---each with distinct predictions about sectoral mobility, geographic portability, and the time path of treatment effects. We test these predictions empirically. Employment gains transfer across sectors but concentrate within local labor markets, indicating portable skills rather than firm- or sector-specific retention. Treatment effects grow rather than attenuate over time, suggesting that human capital accumulation is the primary driver of persistence. The program also reduced informal self-employment and shifted employment from services into manufacturing, consistent with the acquisition of transferable skills during training placements. Together, these results indicate that on-the-job training builds durable workplace skills, while the verified employment record helps workers overcome the information barriers that initially excluded them from formal jobs.
Presented by: Nicolás Campos, The University of British Columbia
Job Search Assistance for Vocational Students Abstract The employment rate of vocational students is 28% six months after graduation in France. This paper evaluates \textit{AvenirPro}, a job search assistance and counseling program specifically designed for vocational students. The program consists of two components. First, caseworkers from the French Public Employment Service deliver in-class interventions during the final year of vocational schooling. These sessions aim to equip students with the knowledge and skills necessary to conduct effective job searches and prepare for interviews. Second, the same caseworkers provide individualized support during the five months following graduation. We designed and implemented a large-scale field experiment with randomization at three levels. First, we randomly selected schools. Second, within treated schools, classes were randomly selected to participate in the group-level counseling sessions. Third, among students in treated classes, half were randomly assigned to receive post-graduation support. Our preliminary results indicate a substantial increase in employment probability of at least 25% six months after graduation among treated students. This effect is driven by the first part of the program.
Presented by: Jeremy Hervelin, ZEW
Restricting Temporary Contracts Increases Firm-Provided Training: Evidence from Spain Abstract We study whether a shift from temporary to permanent contracts strengthens firms’ incentives to invest in worker training, exploiting Spain’s 2022 labour market reform, which sharply restricted the use of temporary contracts. Using 3.1 million online job postings from 2018 to 2024, we implement a difference-in-differences design that leverages pre-reform variation across occupations in their reliance on temporary hiring. Occupations more exposed to the reform experienced a large shift from temporary to permanent contracts and, simultaneously, a significant increase in advertised firm-provided training relative to less exposed occupations. Training in job postings rose by a relative 4.3 percentage points in the most exposed occupations, fully closing the pre-existing gap between occupations with high and low temporary employment. These results provide evidence that longer expected employment duration increases firms’ intended investment in worker training, highlighting an important channel through which labour market regulation can shape human capital formation.
Presented by: Carlos Victoria-Lanzon, Universidad Complutense de Madrid
Session 2: Health June 24, 2026 12:40 to 14:20 Location: Room 2
Session Chair: Omar Fieles-Ahmad, Otto-von-Guericke University Magdeburg
Session type: contributed
Health insurance in Kenya: What do survey diaries tell us? Abstract Expanding access to affordable health insurance is expected to reduce barriers to universal health coverage, yet uptake remains low among poor households across low- and middle-income countries. Most interventions to boost enrollment have achieved only modest results, even with substantial premium subsidies. We exploit the randomized roll-out of Kenya’s National Health Insurance Fund and unique granular data collected weekly over 18 months to examine individual and household responses to incentives for joining a public health insurance scheme. We find that combining a 12-month full subsidy with hands-on support from trusted agents to navigate administrative and registration procedures substantially increases enrollment: take-up rises by 54.7 percentage points, reaching about 80% of individuals, equivalent to nearly 95% of average household enrollment in treated villages. This leads to lower out-of-pocket health spending, faster and more frequent care-seeking when ill or injured, and reduced severity of health problems. These results contrast with many existing schemes that improve financial protection and, to some extent, utilization, but rarely show consistent health gains, raising doubts about whether financial protection alone justifies expanding subsidized insurance. Our findings demonstrate that health insurance does improve health outcomes in low- and middle-income countries, strengthening the case for its expansion. Methodologically, they highlight the value of high-frequency data collection throughout the study period to detect impacts when present.
Presented by: Mahe Clotilde, Vrije Universiteit Amsterdam
Market Impacts of Removing Prescription Requirement for Emergency Contraception in Italy Abstract Italy removed the mandatory prescription requirement to access pharmacological emergency contraception for women over 18 years of age with a staggered approach, at first in May 2015 for the brand EllaOne, and subsequently for the brands Norlevo and Escapelle, with the final reform introduced in March 2016. Using national monthly data on pharmacy stock from 2013 to 2017, I study the impact of these regulatory changes on the sales and revenues of hormonal contraception categories and emergency contraception brands. I implement segmented regression specifications to analyze the outcomes both at the aggregate and at the brand level. Results for emergency contraception find a significant growth of 18% in sales and 64% in revenues following the May 2015 reform, mainly driven by EllaOne, which captured a large portion of the market after becoming prescription-free and caused Norlevo and Escapelle to experience declines. However, after March 2016, there is evidence of a new market shift, with Norlevo sales and revenues rebounding while EllaOne saw a modest reduction. Still, easier access to emergency contraception caused minimal substitution away from other hormonal methods deemed safer. These dynamics suggest that reducing regulatory barriers can boost pharmacological market performance and empower individuals in their reproductive choices.
Presented by: Camilla Genitoni, Università degli Studi di Milano
Gender Differences in Healthcare Utilisation - Evidence from Unexpected Adverse Health Shocks Abstract This paper provides causal evidence on gender differences in healthcare utilisation, challenging common assumptions in the male–female health–survival paradox, whereby women live longer but experience higher morbidity. Using comprehensive Danish administrative healthcare data, we exploit plausibly random variation in treatment timing to identify the causal impact of adverse health shocks, such as first-time non-fatal heart attacks or strokes, on healthcare use. We find that following such shocks, men consistently increase healthcare utilisation more than women. Our results also highlight mechanisms underlying these gender disparities in health outcomes and contribute to policy discussions on healthcare equity and targeted interventions.
Presented by: Seetha Menon, University of Southern Denmark
The Fentanyl Shock: Synthetic Opioids and the Supply of Organ Donors Abstract We study how the diffusion of illicit synthetic opioids affected the supply of deceased organ donors in the United States between 2012 and 2019. As fentanyl spread rapidly from east to west, it produced sharp, uneven increases in overdose mortality across states. We first exploit this diffusion using a state-level instrumental-variables design that interacts longitude with time, measuring fentanyl exposure using mortality and drug seizure data. We then implement a complementary difference-in-differences design at the organ procurement organization (OPO) level that compares regions with early versus delayed fentanyl entry arising from heroin-market segmentation. Both approaches show that fentanyl exposure caused large increases in overdose-driven donor supply and transplanted organs, revealing how a lethal epidemic reshaped the availability of transplantable organs.
Presented by: Omar Fieles-Ahmad, Otto-von-Guericke University Magdeburg
Session 3: Education 1 June 24, 2026 12:40 to 14:20 Location: Room 3
Session Chair: Nina Guyon, Paris School of Economics
Session type: contributed
When Educational Investment Fails to Pay in Labor Market Abstract This study examines the causal effect of private senior high school attendance on academic achievement and long-term socioeconomic outcomes in Taiwan using comprehensive panel data from the Taiwan Education Panel Survey (TEPS) that tracks students from lower secondary education through age 31. Employing value-added models with extensive controls for prior achievement and family background, complemented by coarsened exact matching (CEM), inverse probability weighting (IPW), restricted sample analyses conditioning on private junior high attendance, and bounding tests for unobserved heterogeneity, this study documents substantial, robust, and remarkably subject-specific private school effects on standardized test performance. Private school attendance increases the probability of achieving high grades by 8-11 percentage points in Mathematics and Science on the General Scholastic Ability Test (GSAT), with even larger effects of 6-20 percentage points on advanced mathematics subjects in the Advanced Subjects Test (AST). These STEM advantages persist across all identification strategies and exhibit delta values exceeding 2.0 in bounding tests, indicating unobservables would need to be more than twice as important as observables to eliminate the effects. In contrast, private schools demonstrate weak, inconsistent, or null effects in Chinese, English, and Social Studies, suggesting genuine comparative advantages in STEM instruction rather than uniform quality superiority. However, despite these dramatic test score gains, private school attendance shows no significant effects on university graduation rates, full-time employment, management occupation, or high income by ages 30-31, with only suggestive evidence of possible advantages in postgraduate education enrollment. This profound disconnect between short-term academic achievement and long-term economic outcomes suggests that private school STEM advantages function primarily as positional goods in zero-sum competition for university admission rather than as investments generating lasting productivity gains. The findings have important implications for understanding educational stratification in Taiwan's highly tracked system and raise concerns about reverse redistribution when relatively lower SES families pay tuition for advantages that do not translate into improved labor market outcomes, contributing to credential inflation without collective human capital gains.
Presented by: Chia Nian Lee, National Tsing Hua University
One-way Egalitarianism: Shocks to First-borns and Sibling Spillovers Abstract How do families adjust educational investments when one child is hit by an adverse shock? Using longitudinal data from Indonesia and detailed drought exposure histories, we show that extreme drought at school entry reduces completed schooling by about 0.4 years and lowers adult consumption by 3%. Beyond these individual effects, we analyze how early shocks reshape the intrahousehold allocation of education and alter the distribution of human capital within families. First, shock-affected nonfirstborns experience a decline in their relative educational standing within the sibling group, widening within-family inequality rather than triggering compensatory reallocation. Second, shocks to firstborn children reduce the educational attainment of younger siblings, including those not yet conceived at the time of the shock. These negative spillovers are concentrated among same-gender siblings, particularly younger brothers of male firstborns. We find no evidence that the effects operate through fertility adjustments or reduced financial and non-financial support of the firstborn. Instead, the results point to persistent shifts in parental investment strategies that propagate early disadvantage across birth cohorts within families.
Presented by: Sarah Leuthold, University of Amsterdam
Harnessing Genetic Variants for Local Average Treatment Effect Estimation Abstract When multiple instruments are available, conventional instrumental variable estimators aggregate across potentially heterogeneous margins of compliance and may yield effects that lack a clear economic interpretation. The problem is compounded when some instruments violate the exclusion restriction, as is common in certain empirical contexts such as those using genetic variants as instrumental variables. We propose a clustering-based plurality framework for instrumental variable estimation that jointly addresses instrument heterogeneity and invalid instruments. Rather than imposing a single common causal parameter, our approach groups instruments according to similarity in first-stage and applies a plurality rule on subgroups with similar reduced-form to identify locally valid subsets. This yields a collection of margin-specific local average treatment effects instead of a single pooled estimate. We extend plurality-based identification to settings with non-mutually exclusive instruments, such as Mendelian Randomization designs where all individuals are exposed to all genetic variants. We illustrate the method in a two-sample Mendelian Randomization analysis of the causal effect of educational attainment on smoking participation. Our results confirm a negative causal effect of education on smoking that remains robust under pleiotropy-robust estimators, while revealing substantial heterogeneity across instrument-defined margins that is masked by pooled IV approaches. The framework provides a unified way to interpret and validate high-dimensional instruments in the presence of both treatment effect heterogeneity and potential violations of exclusion.
Presented by: Martin Huber, University of Fribourg
The Effects of Violence at School: Evidence from a Large Scale Randomized Experiment Abstract Using a large scale randomized experiment, we evaluate the causal impact of a social mediation program at school on peer violence as well as on students' socio-emotional and educational outcomes. The intervention consists in professional mediators who handle violence issues in socially disadvantaged junior high schools and we follow students up to the end of high school. The intervention has no average treatment effect on the whole set of students. However, we find substantial treatment heterogeneity with 10 percent of students experiencing a strong decrease in victimization, be it defined as violence in general or repeated violence i.e., bullying. These students who benefit from the program are those who are more often victim of violence in the control group. They also tend to be more often boys, younger, and socially more disadvantaged. For these students, the reduction in violence is accompanied by significant strong improvements in an index of socioemotional outcomes as well as in an index of academic achievement over junior high and high school including performance and the type of final high school diploma. The intervention thus closes 81\% of the gap in academic achievement observed in the control group between these students who would benefit from the intervention and other students.
Presented by: Nina Guyon, Paris School of Economics
Session 4: Regional growth & transport June 24, 2026 12:40 to 14:20 Location: Room 4
Session Chair: Giovanni Rodríguez-Fernández, Complutense University of Madrid
Session type: contributed
On Track to New Jobs: High-Speed Rail and Worker Mobility in Italy Abstract Using Italian administrative data on the population of private-sector workers from 2005 to 2022, we examine how the opening of high-speed rail (HSR) routes expands labor market opportunities. To this end, we estimate gravity-type models at a high level of spatial granularity (municipality) and implement a staggered difference-in-differences research design. We find that the development of new HSR routes induced workers to change their employment location along these routes, with heterogeneous treatment ef- fects by gender, age, contract type, and job qualification. Moreover, such effect is found to be increasing over time, eventually generating additional job relocations of about 12,000 workers per year, and largely due to flows between large municipalities and from small to large ones. The estimation of an AKM model suggests that the connection to the HSR network induced an increase in assortative matching at municipal level.
Presented by: Maurizio Conti, Università di Genova
Brain Drain on the Fast Train: High-Speed Connections and Student Mobility Abstract This paper studies how improved transport connectivity shapes the spatial distribution of prospective human capital, exploiting the staggered expansion of Italy’s high‑speed train (HST) network. Using administrative data on university enrollments and a difference‑in‑differences design, we estimate the causal effect of gaining HST access on student mobility. The opening of a high‑speed connection leads to a persistent rise in student outflows from treated provinces, with no corresponding increase in inflows. This asymmetric response, concentrated in lagging regions and driven by long-distance moves toward major urban hubs, suggests that infrastructure-led integration may exacerbate regional disparities in human capital allocation.
Presented by: Edoardo Frattola, Banca d'Italia
Use It and Lose It: Short- and Long-Term Effects of Transport Subsidies Abstract This paper provides causal evidence on the short- and long-term effects of a 73% fare subsidy for young adults in Madrid, exploiting an age-25 eligibility cutoff via regression discontinuity. The subsidy increases public transport use by 6.3 percentage points and reduces car travel by 6.5 percentage points while active, but effects decay to 4.5 percentage points one year after removal and vanish thereafter for most users. Persistence emerges only among price- or congestion-sensitive prior transit users (3.3 percentage points). These findings indicate that habit formation requires pre-existing behavioral foundations, highlighting the limits of temporary incentives and providing actionable insights for designing targeted and sustainable urban mobility policies.
Presented by: Giovanni Rodríguez-Fernández, Complutense University of Madrid
Session 5: Labour markets June 24, 2026 14:50 to 16:30 Location: Room 4
Session Chair: Andreu Arenas, University of Barcelona
Session type: contributed
Why Do Firms Use Fixed-Term Contracts? Abstract This paper examines why Italian firms rely on fixed-term contracts (FTCs) and quantifies the relative importance of seasonal needs, worker screening, and buffer-stock motives. Using matched employer–employee data for 2013–2017 merged with firm-level balance-sheet information, we show that two ex-ante contract features — the seasonal label and the initial duration — are highly informative of firms’ hiring intentions. Use of seasonal contracts is limited and concentrated in a narrow set of sectors, while longer initial durations strongly predict conversion into permanent jobs, reflecting screening behavior driven mainly by firm-specific factors. In contrast, most FTCs are short, non-seasonal, and concentrated in firms facing lower productivity and higher revenue volatility, consistent with buffer-stock adjustment. A clustering exercise confirms that about 7 percent of firms predominantly use seasonal contracts, one third use FTCs for screening, and roughly 60 percent rely on them to manage uncertainty. These findings highlight contract duration as a key policy-relevant margin and help shed light on structural factors that may contribute to Italy’s stagnating productivity.
Presented by: Matteo Sartori, Banca d'Italia
Consumption Responses to the Minimum Wage Abstract How much does a minimum-wage hike translate into higher household spending—and can this demand impulse meaningfully affect employment? We provide causal estimates of the consumption response to Spain’s 2019 minimum wage increase by combining Household Budget Survey (HBS) expenditure microdata with Social Security records (MCVL). Our identification relies on an instrumental-variables (2SLS) strategy that instruments cell-level income with a wage-gap exposure shock: the pre-reform wage shortfall to the new statutory minimum (the “wage bill gap”) aggregated by region-sector-gender cells. Using a pseudo-panel design with cell fixed effects and controls, we estimate a marginal propensity to consume (MPC) between 0.74 and 0.86, implying that affected households spend most of the income gains induced by the reform. We then translate the implied aggregate consumption increase into different back of the envelope employment scenarios. The results suggest that the consumption-driven demand expansion could have supported job creation, potentially offsetting part of the employment losses implied by cost-side evaluations. Overall, our findings highlight that minimum-wage policies operate through both supply- and demand-side channels, and that ignoring the latter can misstate their aggregate labor-market effects.
Presented by: Rubén Gonzálvez, Universidad Complutense de Madrid
How Climate Investments Reshape Green Skill Demand in Europe Abstract This paper estimates the causal impact of European climate action investments on labour market dynamics, with a particular focus on the emergence and diffusion of green skills. Using more than 250 million online job advertisements from the Web Intelligence Hub–Online Job Advertisement (WIH-OJA) database across European regions between 2018 and 2024, we construct two measures of green skill demand capturing the intensity of green skill requirements within vacancies and their pervasiveness across postings. We link regional labour demand to climate-related expenditures financed through EU Cohesion Policy during the 2014–2020 programming period and address endogeneity in fund allocation using a shift-share instrumental variable strategy based on national mitigation investment shocks and regional exposure shares. Our results show that higher regional exposure to climate action funding significantly increases both the intensity and diffusion of green skill requirements in online vacancies. These findings indicate that early, place-based public investments in climate action play an important role in shaping firms’ skill needs and accelerating the incorporation of green competencies into hiring practices. The paper contributes to the growing literature on the labour market effects of environmental policy by providing direct evidence on skill demand, and offers relevant insights for policymakers, education systems, and firms seeking to support workforce preparation for the green transition.
Presented by: Giovanni Mellace, University of Southern Denmark
Remote work and productivity: evidence from the judiciary Abstract This paper investigates the productivity effects of working from home (WFH) in the public sector, using administrative panel data and multi-stakeholder surveys from the Judiciary in Catalonia. Exploiting a mandatory return-to-office policy implemented in December 2023, the authors employ a difference-in-differences design to measure the impact of ending WFH arrangements. The findings indicate that WFH increased core administrative output: the return to in-person work caused a 5.6% decline in procedural output per worker, with no corresponding change in technical error rates or absenteeism. However, the study documents a "multitasking trade-off," where gains in observable output (such as filings) came at the expense of unobservable service quality tasks, specifically coordination with lawyers. While internal supervisors favored the focus provided by WFH, external stakeholders reported dissatisfaction with service access, suggesting that hybrid models require adjusted monitoring for hard-to-measure tasks.
Presented by: Andreu Arenas, University of Barcelona
Session 6: Health & aging June 24, 2026 14:50 to 16:30 Location: Room 2
Session Chair: Andrei Tuiu, University of Amsterdam
Session type: contributed
Learning from the Oldest: The Impact of Ageing on Healthcare Expenditure Abstract This paper estimates the causal impact of population ageing on healthcare expenditure. Using rich individual healthcare administrative records combined with historical demographic data from Lombardy, Italy´s largest region, we build a highly granular panel covering all municipalities over 2011-2023. To address endogeneity, we exploit predicted variation in ageing across municipalities based on lagged local age structures and regional cohort-specific survival rates. Our instrumental-variable (IV) estimates show that population ageing significantly increases healthcare expenditure per-capita across major domains of care. A 1\% increase in the share of the population aged 60 and above raises per-capita expenditure by 0.69\% for hospitalizations, 1.29\% for pharmaceutical consumption, and 0.47\% for outpatient services. IV estimates are generally smaller than their OLS counterparts, by up to 30--40\% for the oldest age ratio (75+). This indicates that conventional approaches that do not address endogeneity may overstate the true effect of ageing on healthcare expenditure due reverse causality and healthcare-driven migration. Our findings from one of the most advanced country in the demographic transitions, provide credible causal insights into the fiscal pressures to be faced by healthcare systems across ageing societies.
Presented by: Sara Muzzì, University of Milano Bicocca
Retirement age and cardiovascular health: Evidence from Italian administrative data Abstract This paper investigates the causal effect of retirement timing on cardiovascular health among Italian private-sector workers. Using longitudinal administrative data from the Work Histories Italian Panel (WHIP-Health), which combines social security and hospitalization records, we estimate the impact of delaying retirement on the probability of cardiovascular disease (CVD) hospitalization at ages 68–70. To address endogeneity in retirement timing, we adopt an instrumental variable strategy that exploits exogenous variation in retirement age induced by individuals’ month of birth. Our results indicate that postponing retirement by one year increases the probability of CVD hospitalization by 1.7 percentage points (approximately 24% of the sample mean). The adverse health effect is concentrated among men with poorer baseline health, in manual or physically demanding occupations, and those with lower education and earnings. For women, estimated effects are smaller and statistically imprecise, likely reflecting distinct labour market trajectories and selective retirement patterns. These findings contribute to the international debate on the health consequences of delayed retirement in ageing societies with rising statutory pension ages. They highlight the unequal health costs of working-life extension and underscore the need for pension policies that account for differences in health status, work intensity, and socioeconomic position across workers.
Presented by: Chiara Ardito, European Commission Joint Research Centre
How raising the full retirement age affects women's early retirement choices: insights from the interaction of two policies Abstract This paper analyzes how a reform increasing statutory retirement age from 60 to 64 affected women’s incentives for early retirement. In Italy, women can anticipate retirement at 57 (with 35 contribution years), but subject to an annuity penalization. Using Italian administrative data, we compare women eligible for the early retirement scheme before and after the reform, finding a small effect on women’s retirement age, but a substantial negative effect on annuity. Effects are stronger for women with low labor market attachment or working full-time, suggesting that reconciliation of paid and unpaid works is an important driver of early retirement choices.
Presented by: Ylenia Brilli, Ca' Foscari University of Venice
What are we waiting for? The labor market and health effects of elective surgery waiting times among the 50+ population Abstract Long waiting times for elective surgeries are a persistent limitation of universal healthcare systems across Europe, an issue intensified by ageing populations and the increasing burden of age-related conditions. Delaying life-improving procedures may lead to detachment from the labor market, an increased reliance on the social safety net and a loss in well-being generally. Furthermore, it risks generating financial spillovers onto the pension and welfare systems. This paper leverages rich administrative data on all specialist medical procedures in the Netherlands to study the costs of longer waiting times for hip replacement surgeries in osteoarthritis patients aged 50 and older. I first study the labor market, health and welfare consequences of osteoarthritis onset in a staggered difference-in-differences setting. I then use an instrumental variables design leveraging plausibly exogenous variation in provider-level waiting times to study the costs of delays, stratifying the sample by likely retirement status. Finally, I study the heterogenous consequences of delaying treatments by gender and education level. The results will provide valuable insights into the direct and spillover effects of longer waiting times for elective procedures.
Presented by: Andrei Tuiu, University of Amsterdam
Session 7: Higher education 1 June 24, 2026 14:50 to 16:30 Location: Room 3
Session Chair: Maxime François, KU Leuven
Session type: contributed
Backward Spillovers and Equalized Access: The Effects of the Bologna Process across Italy’s Education System Abstract This paper estimates the causal effects of the Bologna Process reform in Italy, analyzing the impact of the introduction of the "3+2" degree structure at different stages of the educational system. We implement multiple Difference-in-Differences designs on a nationally representative dataset composed of more than 140,000 high school graduates and 220,000 university graduates. We find that the reform increased graduations from the academic-oriented high school track by +5.50 p.p., revealing a "backward spillover effect" on high school completion. For university-related outcomes, we estimate a positive effect on enrollments (+5.53 p.p.) and in-time graduations (+7.22 p.p.) and a negative effect on dropout (--2.70 p.p.). The effects are always larger for students coming from the most disadvantaged parental backgrounds, particularly when combined with high ability. These findings suggest that the Bologna Process was effective in reducing barriers to educational investment, especially for talented students with financial constraints, thus challenging the narrative of the reform attracting low ability students. Finally, descriptive labor-market analyses show a convergence of performances for graduates with different parental backgrounds, but only among Bachelor’s graduates.
Presented by: Marco Morelli, University of Luxembourg
Second Chances in Higher Education: Turning Setbacks into New Opportunities Abstract Using administrative data from France’s centralized college admissions platform, I analyze students already enrolled in higher education who apply to switch to a different program, typically restarting from the first year. To estimate the effects of postsecondary redirection, I exploit a fuzzy regression discontinuity design based on the lowest admission cutoff among the programs to which each student applied. For applicants just above the cutoff, admission to a new program increases matriculation and graduation rates by 38 and 23 percentage points, respectively, measured within six years of the new application. The effects are strongest for students shifting to two-year vocational programs. Although postsecondary redirection entails additional public spending, higher graduation rates offset these costs, making it a cost-neutral way to correct early misalignments in students’ educational paths.
Presented by: Nagui Bechichi,
Unrealistic Educational Aspirations: Behavioral Insights From Parental Educational Decisions in Albania Abstract This paper documents substantial parental misperceptions about academic trajectories and college eligibility at the time of secondary school track choice. Using administrative data combined with a large-scale survey of parents in Albania, we show that parents systematically overestimate their child’s future academic performance and chances of meeting college eligibility thresholds, while being more pessimistic about similarly performing peers. These misperceptions coexist with high enrolment in academic high schools, even among students who are unlikely to be college-eligible. In contrast, parents beliefs about vocational education returns and stigma appear broadly accurate. The findings highlight the potential role of information frictions in shaping secondary school track choices.
Presented by: Theodor Kouro, Charles University & Economics Institute of the Czech Academy of Sciences
When University Goes Global: The Causal Impact of International Peers on Domestic Students' Performance Abstract Student mobility offers a unique opportunity for students to discover, exchange and develop themselves. Accordingly, the share of international students in tertiary education has been steadily increasing, alongside the growing prevalence of courses and programs taught in English, which facilitates mobility for domestic students. This paper contributes to the literature by causally and separately evaluating these two intertwined phenomena. Using administrative data from a top-50 comprehensive European university, we examine (i) the switch of courses to English as a Medium of Instruction (EMI) and (ii) the subsequent increase in the share of international students enrolled in these courses. We rely on recent advances in staggered Difference-in-Differences designs with two treatments and a continuous treatment variable. Our results indicate that an EMI switch leads to a 27.5% increase in the share of international peers relative to the mean, followed by stabilization over time. For domestic students, the EMI switch initially entails adverse short-run effects, namely a 2.5% increase in retake exams and no significant change in grades, before adjusting in subsequent years. An increasing share of international peers generates small but positive effects on domestic academic performance, corresponding to a 0.14-point increase on a 20-point grading scale and a 1.1% reduction in retake exams. Overall, these findings underscore the benefits of greater internationalization from its positive academic effects, a conclusion that contrasts with the current political trend toward restricting English-taught programs in higher education.
Presented by: Maxime François, KU Leuven
Session 8: Labour 1 June 24, 2026 14:50 to 16:30 Location: Room 1
Session Chair: Giulia Tarullo, RWI-Leibniz Institute for Economic Research
Session type: contributed
The Effect of Subsidies on Training: Evidence from the United States Abstract Labour demand is rapidly evolving due to fast-paced technological change (Lightcast, 2022). The speed of this evolution causes the demand for skilled workers to outpace supply in such sectors (Feng, 2021). Consequently, employers in the United States lack qualified workforce in multiple industries (Lerman et al., 2019, Magnini et al., 2024). One solution to mitigate labour market shortages is non-formal Registered Apprenticeship programmes, which quickly and adaptably impart work-relevant skills to the workforce (Lightcast, 2022, OECD/ILO, 2017, Kuczera, 2017). Participation in non-formal training programmes among 25–64-year-olds in OECD countries stands at 40% (Denzler et al., 2025). Widespread participation in such programmes prevents human capital depreciation and furthers lifelong learning, which is particularly important in aging OECD economies. Amid rising automation of routine tasks, participation in non-formal training programmes also allows workers to subsequently be more involved in non-routine tasks (Tamm, 2018), relatively more shielded from automation. Furthermore, non-formal training is essential to prepare the United States workforce for the implementation of large-scale investments in cutting-edge technologies and infrastructure investments, such as the CHIPS and Science Act, or the Infrastructure Investment and Jobs Act (United States Government Accountability Office, 2025). The subsidisation policy evaluated in this paper is the American Apprenticeship Initiative (AAI). AAI subsidies totalled $154M (USA Spending, 2015, see Appendix Table 1). These $154M are part of the $1B spent by the federal Department of Labor between 2015 and 2022 as part of vocational training subsidies (Butrica et al., 2023). The AAI aims to increase relatively low employer engagement and upskill the United States workforce through Registered Apprenticeship (United States Department of Labor, 2015, OECD/ILO, 2017, Lerman et al., 2019). The AAI grant period lasted five years; October 1st, 2015, to September 30th, 2020 (United States Department of Labor, 2015). AAI grant funds were distributed to grantees operating in 25 states and Washington D.C. Grantees were either state governments, nonprofit organisations, public/state-controlled higher education institutions (e.g. technical or community colleges), city or township governments, or workforce development agencies (USA Spending, 2015). Grantees are listed in Appendix Table 1. Within their respective operating state, i.e. their place of performance, AAI grantees were to fund Registered Apprenticeships in one of the four manners elicited in Appendix Table 2. Chiefly, these channels were reimbursing employers for OJT costs, and RTI providers for Registered Apprentice’s tuition fees (United States Department of Labor, 2015). I provide more details on the AAI’s implementation in the background section of this paper (Section 2). The objective of the AAI was to fund the registration of 34,000 new Registered Apprentices over five years (National Governor’s Association, 2020). This is 8% of active Registered Apprentices in 2015 (ApprenticeshipUSA, 2024). In addition, federal Department of Labor guidelines dictated that AAI grantees must channel funding towards industries that were most reliant on H1B visas, i.e. foreign workforce. The AAI was thus to develop the United States’ workforce skills within these critical industries to lessen the dependent on foreign labour (United States Department of Labor, 2015). The United States Citizenship and Immigration Services (2025) indicate that on average, between 2009 and 2015, NAICS industries in which employers used the largest fraction of total H1B visas were Professional and Scientific Services (58% of total initial approvals of H1B visas), Educational Services (12%), Advanced Manufacturing (7%), Healthcare and Social Assistance (5%), and Information (4%). The United States Department of Labor (2015) required AAI grantees to, within the states they operate, finance Registered Apprentices and their programmes active in these industries. This paper’s empirical analysis uses federal administrative data on start date, state, county, occupation and industry of Registered Apprentices from RAPIDS (Registered Apprenticeship Partners Information Database System, ApprenticeshipUSA, 2024). Employers across the United States must submit individual Registered Apprentice record data to the Department of Labor’s Office of Apprenticeship, which is then recorded into the nationally representative RAPIDS database (Bruno and Manzo, 2025). This paper uses two identification strategies. In the empirical methodology section, I detail how each identification strategy overcomes selection into treatment; AAI funds were not distributed randomly. The first is triple difference. It considers state, time and industry variation in treatment. The second leverages spatial and time variation in treatment in the form of a spatial difference-in-discontinuity. I use variation across counties located in states of opposite treatment status and variation in AAI treatment timing. The use of the latter identification strategy follows its reliance on alternative identification assumptions, and the findings of Giroud et al. (2024) and Belenzon and Schankerman (2013). Spread of information and knowledge-sharing are very localised, and decays with distance rapidly. Furthermore, knowledge-sharing is stronger within same-industry geography clusters, which may occur within narrow bandwidths about the border (Kim, 2023). Consequently, awareness about the AAI and familiarity with Registered Apprenticeship programmes likely negatively depends on distance and is more similar within a narrow bandwidth about state borders. It thus follows to model the outcome variable as a function of distance to the border of a state of opposite treatment status. Both identification strategies yield aligned results. The AAI has not significantly increased the number of new Registered Apprentices. Robustness checks corroborate these baseline results. Additional difference-in-discontinuity estimates indicate that the effect of the AAI is not significantly stronger for treated industries located within treated counties in a narrow bandwidth about the state border. If the triple difference estimates were significant, I would find that $1M spent through the AAI led to 22 Registered Apprentices hired in treated states, industries, and over the first year of the AAI-treatment period. This equates to $45,454 per job-year in treated states and industries. Hiring 22 Registered Apprentices at the prevailing competitive wage rate for Registered Apprentices in the relevant industries, states and time period ($40,114) would however cost $882,508 annually, therefore would be more economical. Nevertheless, this cost-benefit analysis has a caveat. Washington D.C., New York, Massachusetts, Oregon, Washington, and Wisconsin do not make data available in RAPIDS relative to industry or occupation. These states were thus excluded from triple difference analysis and consequently omitted from this cost-benefit analysis.
Presented by: Guillaume Morlet, ETH Zurich
Causal effect of subsidized employment on labour market attachment outcomes: older workers case in population ageing economy Abstract In this paper, I evaluate the effect of subsidized employment for people over the age of 50 in Poland's labor market. Data from the Polish employment register provide the foundation for a detailed analysis through propensity score matching and regression discontinuity design. The analysis focuses on the program's impact on the return to the unemployment register and the duration of assumed sustained employment. Results from PSM indicate that subsidies generally increase the likelihood of returning to the unemployment register by 12\%, 16\%, and 17\% within two, three, and four years, respectively. The effect is particularly strong for females. Conversely, subsidies show a relative advantage over other active labor market policies. Regression discontinuity analysis reveals that extending subsidy protection from 18 to 36 months at age 60 does not effectively prolong employment duration. The study's outcomes cast doubt on the long-term benefits of the program and highlight the potential for policy refinement to address the specific needs of an aging workforce.
Presented by: Aleksandra Sojka,
Payroll tax incidence for hiring the first employee Abstract This paper studies payroll tax incidence at firm entry, a setting rarely examined in the literature. We exploit a 2016 Belgian reform that permanently exempted new employers from payroll taxes for their first employee, sharply reducing hiring costs and encouraging entry of lower-productivity firms. Using administrative data and a cohort-based difference-in-differences design, we find that the reform reduced payroll taxes while raising both wage rates and working time. To interpret these results, we develop a unified model linking firm entry, firm productivity, and tax incidence. We then propose a theory-guided trimming method to address compositional changes of firm entry. The yielded estimates remain close to the baseline results, suggesting productivity differences matter little for entry incidence. Employees capture about one-third of the payroll tax relief on a per-FTE basis, which suggests the importance of wage bargaining at firm entry.
Presented by: Haotian Deng,
Experience Rating in Short-Time Work Abstract Short-time work (STW) has been a key labour market policy response to both the Great Recession and the COVID-19 crisis, subsidising temporary reductions in working hours to prevent layoffs. However, by funding reduced working time, STW inherently incentivises excessive take-up absent financial incentives. This paper examines two Belgian STW system reforms—in 2005 and 2012—that introduced taxes for firms based on their employees’ STW intensity to mitigate distortions in working hours, covering about 13% of firms in STW. Leveraging employer-employee longitudinal data and bunching designs, we provide novel evidence that both reforms significantly reduced STW intensity per worker, with strongest effects observed among higher-wage jobs. Additionally, we identify short-lived yet notable spillover effects at the firm level: after the reform, firms adjusted their STW usage, distributing reductions in working time more evenly across employees.
Presented by: Giulia Tarullo, RWI-Leibniz Institute for Economic Research
Session 9: Pay transparency & female labour supply June 25, 2026 9:00 to 10:40 Location: Room 1
Session Chair: Anna Cecilia Rosso, University of Milan
Session type: contributed
Measuring and Decomposing the Gender Gap in Wage Negotiations: Evidence from Expatriates in the Netherlands Abstract This study quantifies the contribution of gender differences in wage negotiations to the gender wage gap. To isolate the importance of wage negotiations for earnings, we exploit a substantial negative income shock among mid- and high-earning migrants in the Netherlands once their tax benefits (the “30%-rule”) expire. Expiration occurs after 5 to 10 years since arrival, and reduces net income often by at least €10,000 annually, generating a strong incentive to renegotiate the wage. We estimate that men experience an average wage increase that is 0.6–1.5 percentage points higher than that of women. The difference is particularly pronounced among higher earners. Using an Oaxaca-Blinder decomposition applied to the wage increase, we find that the negotiation gap is explained by gender differences in income levels, marital status, and industry, and to a smaller extent by country of origin, collective bargaining coverage, firm size, and homeownership; once we account for these factors, the remaining unexplained negotiation gap becomes negligible. This points to two key mechanisms: i) wage inequality: those with higher wages are able to negotiate larger wage jumps, and ii) mobility constraints: family-related ties limit outside options and thus bargaining position.
Presented by: Mark van der Meijden, Vrije Universiteit Amsterdam
Child-Related Transfers, Fertility, and Maternal Labor Supply: Evidence from Reform-Induced Income Shocks Abstract This paper examines the effects of increases in child-related cash transfers on fertility and maternal labor supply using population-wide administrative data and a design that exploits heterogeneous changes in benefit generosity. By comparing outcomes across adjacent treatment-intensity groups, we identify local causal effects in a setting without untreated units. We find that higher transfers increase the probability of having an additional child only over a limited range of benefit increases, with effects concentrated in the middle of the income distribution. At the same time, increased benefit generosity reduces maternal labor supply in the medium run for specific groups. These labor supply responses operate through both employment exits and reductions in working hours and persist even when fertility-related adjustments are shut down. Translating these effects into elasticities reveals substantial heterogeneity across the income distribution, underscoring the importance of local responses and adjustment margins. The findings have implications for the design and evaluation of family policies. Child-related cash transfers can stimulate fertility among financially constrained households, but they may also reduce maternal labor supply along margins that are sensitive to non-labor income. Policies aimed at addressing low fertility and low female employment should, therefore, account for heterogeneous responses and potential trade-offs across margins of adjustment, rather than relying on average effects alone.
Presented by: Chiara Dal Bianco, Università di Padova
Parental leaves and female labour force participation in Italy: the effect of the 2015 reform Abstract This paper investigates the impact of Italy’s 2015 parental leave reform on female labour force participation and employment. The reform extended eligibility for paid parental leave from a child’s third to sixth birthday, potentially improving mothers’ ability to reconcile work and care responsibilities during early childhood. Using quarterly Labour Force Survey data and a Difference-in-Differences (DiD) approach, we compare outcomes for mothers with their youngest child aged 3–5 (treated group) to those with children aged 6–10 (control group) before and after the reform. To address potential confounders, we implement an intention-to-treat design and control for key socio-demographic characteristics. We also perform a propensity score matching combined with the DiD approach to account for some differences in the pre-reform data. Our results indicate that the reform led to a significant increase in labour force participation and employment among treated mothers, particularly in the short term. These findings suggest that expanding access to parental leave entitlements can strengthen women’s labour market attachment.
Presented by: Emma Fantappiè,
Should I stay or should I go? The impact of pay gap transparency on female workers Abstract This paper examines the effects of the UK’s 2017 gender pay transparency policy, which requires large firms to disclose gender pay gap data. Using a novel employer-employee dataset and a regression discontinuity design around the treatment threshold, we find that the policy led to increased job mobility among women, especially high-skilled, native-born workers in non-managerial roles. Treated firms also reduced female hiring, while promotion rates remained unchanged. Our findings suggest that transparency influences workforce composition, contributing to a narrowing of the gender pay gap—primarily through mobility of women.
Presented by: Anna Cecilia Rosso, University of Milan
Session 10: Education 2 June 25, 2026 9:00 to 10:40 Location: Room 2
Session Chair: Jacob Van Belle, KU Leuven
Session type: contributed
Education Reform in Africa Abstract To address the global learning crisis, targeted remediation and structured pedagogy are often considered “great buys” for education policymakers. But successful small-scale programs can be difficult to scale, so that impacts at scale in government schools are often disappointing. We collaborated with the Ministry of Education in an African country to conduct a prospective evaluation of a large-scale education reform that combines two “great buys”. We implement a matched-paired difference-in-differences design on 276 schools that implemented the reform in 2023 (Cohort 1 schools) and 62 schools that implemented the reform in 2024 (Cohort 2 schools). We follow 184 of the Cohort 1 schools into the second year of the reform, presenting results on the effects of the reform over two years in these schools. We also estimate the impact after one year for the 62 Cohort 2 schools. For the Cohort 1 schools after two years, the program resulted in large increases in student learning of Arabic, French, and mathematics. The overall impact was nearly 0.6 standard deviations across all subjects. Although the impact of the reform remains very large relative to impacts that have been found for education programs in other middle- and low-income countries.
Presented by: Kenza Sabri, J-PAL
Compulsory Schooling in the 21st Century: Short-Term Effects of Finland’s 2020 Act on Compulsory Education Abstract This study examines the short-term effects of Finland’s 2021 reform, which extended compulsory schooling to age 18. Using detailed administrative data, I analyze changes in educational outcomes and early labor market outcomes following the reform. The reform is associated with a 20% reduction in dropout rates between ages 15 and 18, with a sharp and temporary increase in dropout rates once the law no longer binds. On average, students in the first affected cohorts are 1.9% (1.5 percentage points) more likely to graduate from secondary school by the age of 20 and complete 1.2 percentage points more courses relative to pre-reform cohorts. Time spent outside of education, employment, or training is reduced by 0.38 months on average, while annual earnings decrease by 86 euros. Heterogeneity analyses indicate relatively stronger responses among students from low-SES and foreign backgrounds and among males.
Presented by: Juho Junttila, University of Helsinki
Internet Access and Educational Attainment: Evidence from India Abstract The expansion of internet access has been promoted in many developing countries to foster economic growth and equalize opportunities. Yet, evidence on the effects of internet access on educational opportunities remains scarce and inconclusive. Documenting and exploiting a massive shock to internet access across all of India, we provide robust evidence for gender-specific impact of expanding internet access to adolescents in India and find no effect for males but positive effects for females. Further, we show that the zero effect for males masks considerable heterogeneity by parental background as individuals with non-educated parents are negatively affected whereas those of more educated parents profit from the internet access expansion.
Presented by: Fabian Reutzel, European Commission DG Joint Research Center
Reducing the Teaching Penalty: The Impact of an Experience-Based Salary Reform on Second Career Teacher Supply Abstract Second-career teachers (SCTs) are a promising group to accommodate teacher shortages. Yet, transitioning into teaching is often financially unattractive due to a ‘teaching penalty’. Experience-based salary policies that recognise prior professional experience outside of education are often proposed to attract and retain SCTs. However, to date, no causal evidence exists on the effectiveness of such interventions. We address this gap by studying career transitions into teaching and early-career turnover among SCTs. Using administrative databases from Flanders (Belgium), we evaluate the effectiveness of an experience-based salary reform that was introduced for starting SCTs with specified eligible degrees and implemented between September 2020 and September 2023. We estimate treatment effects on (1) enrolment of prospective SCTs into teacher education programmes, (2) job-entry of graduate teachers, and (3) turnover rates one year after job-entry employing modern difference-in-differences and triple-differences estimators, carefully assessing the identification assumptions. In our analyses, we focus on teacher education programmes for secondary education. Findings are mixed. While there is a significant and sizeable increase in the enrolment of SCTs in eligible teacher education programmes, we find no treatment effect on the subsequent job-entry and turnover rates of eligible SCTs, although there are anticipation effects concerning turnover. The short-term effectiveness of the salary reform is limited but it slightly alleviates teacher shortages. We complement impact estimates with cost-effectiveness metrics, which is the budgetary cost per additional teacher.
Presented by: Jacob Van Belle, KU Leuven
Session 11: Causal methods & learning June 25, 2026 9:00 to 10:40 Location: Room 3
Session Chair: Franco Peracchi, EIEF
Session type: contributed
On Conditional and Unconditional Quantile DiD: Identification, Weighting, and Estimation Abstract We study quantile difference-in-differences (DiD) in panel settings and clarify the relationship between conditional and unconditional quantile DiD estimands. We formalize Conditional Quantile DiD (CQDD) and Unconditional Quantile DiD (UCQDD), show that they coincide without covariates but generally diverge with covariates, and provide intuition for why empirical conditional and unconditional procedures can differ even in simple designs. We then analyze two-way fixed-effects (TWFE) implementations based on quantile regression and recentered influence functions. While TWFE identifies the intended quantile DiD contrast in the canonical 2x2 design, it delivers density-weighted mixtures with multiple post-treatment periods and, under staggered adoption, combines cohort--time effects with potentially negative weights, undermining causal interpretation. Motivated by these results, we propose clean quantile DiD estimators for staggered adoption based on valid cohort-time comparisons and transparent convex aggregation. Simulations show that the proposed estimators eliminate the large distortions that can arise from TWFE quantile regressions under heterogeneous effects and staggered timing.
Presented by: Camilo Arias, Universidad de la Sabana
GenAI: Substitute or Complement in Evidence Synthesis. A discrete choice experiment with Italian policymakers Abstract Credible evidence on the causal effects of public policies is increasingly produced by the academic community, and evidence-based policymaking is increasingly demanded by governments, yet the production of high-quality evidence syntheses remains costly, slow, and often poorly aligned with policymakers’ time constraints. Recent advances in generative artificial intelligence (GenAI) promise efficiency gains in evidence synthesis, but their adoption in policy contexts depends on whether policymakers trust and value AI-generated evidence. This paper studies policymakers’ preferences for the use of GenAI in producing scientific evidence syntheses for policy design. We implement a discrete choice experiment with 813 Italian policymakers, eliciting trade-offs over synthesis production method (expert-produced, AI-generated, or hybrid AI-with-expert-review), cost, and delivery time. In a between-subject design, we additionally vary whether the AI model is described as general-purpose or specifically validated for evidence synthesis. We find a strong aversion to fully automated AI-produced syntheses, even when conditional on cost and time savings. By contrast, the hybrid AI–expert approach is systematically preferred over both expert and AI approaches, conditional on cost and delivery time. Explicit validation of the AI model significantly increases acceptance of AI-assisted synthesis, reducing resistance to automation and increasing preferences for hybrid production. Heterogeneity analyses show that these effects vary with attitudes toward AI and science but not with gender or institutional role. Overall, the results indicate that policymakers value AI primarily as a complement to human expertise rather than as a substitute, highlighting the importance of task-specific validation and expert oversight for the legitimate adoption of AI in evidence-based policymaking.
Presented by: Sergiu Burlacu, Bruno Kessler Foundation
Covariate Balancing and the Equivalence of Weighting and Doubly Robust Estimators of Average Treatment Effects Abstract How should researchers adjust for covariates? We show that if the propensity score is estimated using a specific covariate balancing approach, inverse probability weighting (IPW), augmented inverse probability weighting (AIPW), and inverse probability weighted regression adjustment (IPWRA) estimators are numerically equivalent for the average treatment effect (ATE), and likewise for the average treatment effect on the treated (ATT). The resulting weights are inherently normalized, making normalized and unnormalized IPW and AIPW identical. We discuss implications for instrumental variables and difference-in-differences estimators and illustrate with two applications how these numerical equivalences simplify analysis and interpretation.
Presented by: S. Derya Uysal, LMU Munich
A Local Differencing Test for the Credibility of Selection-on-Observables Abstract One of the most common research designs employed for the identification of causal parameters from observational data is selection-on-observables. However, justifying the underlying assumptions can be challenging and the common practice of controlling for a large number of covariates may not always suffice. This paper introduces a local differencing test designed to evaluate the credibility of selection-on-observables within specific applications. Our procedure leverages the possibility of constructing, under selection-on-observables, two consistent and asymptotically normal estimators of the conditional average treatment effect (CATE) function. We propose to locally test the null hypothesis that identification is credible at some prediction point by employing the ratio of the difference between the two estimates evaluated at that prediction point to its standard error. We show that an “oracle” version of this statistic follows an asymptotic standard normal distribution under the null hypothesis when either OLS regressions or honest regression forests are employed for CATE estimation. Our theoretical results are validated by a simulation exercise.
Presented by: Franco Peracchi, EIEF
Session 12: Place-based policies June 25, 2026 9:00 to 10:40 Location: Room 4
Session Chair: Mirko Titze, Halle Institute for Economic Research (IWH)
Session type: contributed
Disaster Risk, Saliency Shocks, and Real Estate Markets: Evidence from the Phlegraean Fields Abstract Hedonic models suggest that real estate markets price natural disaster risk, particularly in the aftermath of catastrophic events. But can salience alone—absent any actual disaster and with an underlying hazard already known and pre-existing—shift risk perception and tolerance enough to affect economic behavior and market outcomes? We answer this question by drawing on a natural experiment provided by the Phlegraean Fields, a large volcanic caldera near Naples, Italy. The area is home to over one million residents and is characterized by bradyseism, a distinctive geological phenomenon involving cyclic ground uplift often accompanied by seismic activity perceptible by the population. Combining granular sub-municipal data with an event-study design, we show that the bradyseismic crisis that began in 2023 significantly affected local real estate markets, despite minor physical damage and no abrupt change in the long-known risks of living in the area. We find substantial heterogeneity both between and within affected zones, and the results suggest that the crisis is exacerbating pre-existing housing inequality in the most vulnerable areas.
Presented by: Samuel Nocito, Sapienza University of Rome
How do Firm-Level Capital Subsidies Affect Workers? Evidence from a German Place-Based Policy Abstract In this paper, we study the effects of a German place-based policy that subsidizes firms’ capital investment on the long-term outcomes of workers employed by these firms. Our research design leverages EU-wide rules governing program parameters at the regional level, which generates spatial discontinuities in program eligibility and generosity. Using the universe of matched employer-employee data linked to detailed administrative data on the subsidy program, we estimate dynamic causal treatment effects of the investment subsidies on the labor market outcomes of affected workers over a period of up to 15 years. We also investigate the channels behind these effects by estimating a model in which firms produce goods using capital and heterogeneous labor and face upward-sloping labor supply curves. We find that incumbent workers employed by a firm at the time the subsidy was received earn higher wages, are more likely to be employed, and are less likely to change employers than matched control workers in unsubsidized firms. Importantly, these benefits are highly persistent over time. Workers hired through subsidy-induced recruitment efforts also experience wage and job stability benefits, though the effects are smaller than those for incumbent workers. Our heterogeneity analysis indicates that older workers experience the largest wage gains and remain employed by delaying retirement. The model estimates imply that firms pay higher wages to retain older workers because they are imperfect substitutes in production with younger workers, making it challenging to increase production solely by increasing the number of younger workers. We also estimate an elasticity of substitution between capital and labor of about 1.7, which limits the ability of firms to replace labor with capital in response to investment subsidies. Our estimates imply that the scale effect dominates the capital-labor substitution effect, leading firms to substantially increase employment in response to investment subsidies.
Presented by: Philipp Grunau, Institute for Employment Research (IAB)
Quantitative evaluation of place-based policy-mixes: Lessons from a French regional innovation policy Abstract This paper evaluates the effects of R&D place-based policies implemented by regional governments. We constructed a unique database that provides detailed information on direct collaborative R&D support from France’s leading Auvergne-Rhône-Alpes Region from 2009 to 2021, while controlling for national direct and indirect R&D support policies. Using a staggered event study design, we assess the dynamic policy’s impact on firm-level employment, turnover, exports, and financial performance. The econometric results reveal that innovative firms experience a gradual increase in turnover and employment after receiving regional support, with a more pronounced effect emerging approximately eight years later, supporting the idea that evaluations of innovation policies must be part of a long-term perspective. Additionally, our findings highlight interaction effects between different instruments and government levels. The contribution of regional policies would be particularly significant for companies that do not yet receive national aid.
Presented by: Corinne Autant-Bernard, University of Saint-Etienne
Place-Based State Aid for the Green Transition: Evidence from Germany’s Coal Phase-Out Abstract We study the early labor market effects of Germany’s Structural Strengthening Act for Mining Regions, a large place-based program designed to cushion the regional adjustment costs of the coal phase-out. The Act provides grants in an amount of e 41.09 billion until 2038 and targets a limited set of coal-exposed districts through a mix of infrastructure, innovation, and regional development measures. Using quarterly administrative labor market statistics from the Federal Employment Agency at the district level, we evaluate whether eligible regions exhibit differential trends in employment, unemployment, and job vacancies. Our empirical design applies an event-study framework comparing the districts eligible under the Act to a control group of structurally similar districts. To strengthen pre-treatment balance, we complement two-way fixed effects event studies with an event-study variant of synthetic difference-in-differences. Across all specifications, we find no measurable effects on the three labor market outcomes up to the end of 2024. This pattern is consistent with the program’s early implementation stage: by December 2024, only about e 2.2 billion (5.5%) of total funds had been disbursed, and many major projects are likely to affect labor markets only with substantial lags. We interpret our results as evidence on short-run, eligibility-based impacts in a complex environment shaped by the coal exit and concurrent macro shocks.
Presented by: Mirko Titze, Halle Institute for Economic Research (IWH)
Session 13: Development June 25, 2026 11:10 to 12:25 Location: Room 1
Session Chair: Jochen Kluve, Humboldt-Universität zu Berlin
Session type: contributed
Average and Distributional Impact of Novissi in Togo: Evidence on Wealth, Consumption and Subjective Poverty Abstract This paper evaluates the impact of GD-Novissi, an unconditional cash transfer program implemented in Togo during the COVID-19 pandemic, on household wealth, consumption, and subjective wellbeing. Exploiting panel data from two waves of the Enquête Harmonisée sur les Conditions de Vie des Ménages, we apply a doubly robust difference-in-differences estimator (Sant’Anna and Zhao, 2020) with covariates selected via random forest to address the endogenous, poverty-targeted nature of the program. We find large and significant positive effects on household asset wealth, while effects on per capita consumption and subjective economic difficulty are small and statistically inconclusive at the household level. Quantile regressions reveal that asset gains are concentrated at the lower end of the distribution. Treatment Effect Curves (Verme, 2010, 2026) further suggest that consumption effects are positive for initially poorer house holds and offset by negative effects at the upper tail. An individual-level replication additionally detects a significant negative effect on subjective economic difficulty.
Presented by: Virginia Fidenti, University of Bologna
Long-term health and human capital effects of massive investments in public health and education: Evidence from Cuba Abstract This paper estimates the long-run effects of universal public health provision on health and human capital. We study Cuba’s National Health Service, introduced alongside a nationwide literacy campaign, using newly assembled historical data and synthetic control methods. Relative to a carefully constructed counterfactual, Cuba experienced a persistent 20–30 percent decline in infant mortality and a 1.5-2 year increase in educational attainment. These effects are robust across modern panel estimators. In contrast, life-expectancy gains are sensitive to specification and attenuate over time. Overall, large-scale public investments in health and education generate durable improvements in early-life survival and human capital.
Presented by: Rok Spruk, University of Ljubljana
Linking KfW-Financed Aid Projects and FDI: Global Evidence at the Sub-national Level Abstract This study analyzes the empirical link between German bilateral development finance and Foreign Direct Investment (FDI) at the sub-national (ADM2) level. We construct a unique dataset by merging geo-referenced development aid projects--- implemented by KfW Development Bank---over more than two decades with FDI project data. The analysis investigates four research hypotheses, and finds that: (i) development finance activity is positively and significantly associated with FDI inflows; (ii) the positive link is similarly pronounced in both hard and soft sectors; as well as (iii) irrespective of recipient countries' income level; and (iv) the positive aid–FDI association appears to be driven by projects with stronger implementation, as measured by higher ex-post evaluation ratings. For these higher-rated projects, KfW aid is significantly more likely to be associated with FDI from Germany and the EU. Our findings suggest that FDI may be an important channel through which development aid simultaneously benefits both recipient---by providing capital and technology---and donor countries, by signaling investment opportunities for its enterprises.
Presented by: Jochen Kluve, Humboldt-Universität zu Berlin
Session 14: Firms June 25, 2026 11:10 to 12:25 Location: Room 2
Session Chair: Alexsandros Cavgias Martins Frag, Ghent University
Session type: contributed
American Indian Casino Operations and Adjacent Firm Long-Run Employment and Sales Abstract American Indian reservations face long-standing barriers to private-sector development, making IGRA-enabled tribal casino gaming one of the most consequential sovereignty-driven, place-based interventions in Native communities—yet credible evidence on broader local business spillovers remains limited. We address this gap by linking geocoded National Establishment Time Series (NETS) microdata to federally recognized reservation boundaries and a hand-collected panel of tribal casino opening years from 1990 to 2019, and we estimate dynamic impacts using a staggered-adoption difference-in-differences design (group-time average treatment effects) that compares treated reservations to never-treated reservations, tests for differential pre-trends, and is robust to covariate adjustment. Casino openings generate economically meaningful and persistent growth in reservation business activity: average establishment sales increase by 47% and employment increases by 50% after opening. Importantly, incumbent establishments also expand, with sales increasing by 17% and employment increasing by 26%, indicating that development benefits extend beyond new entry and beyond the casino itself, consistent with local demand, business-to-business spending, and visitor-spending spillovers. When excluding the most gaming-adjacent industries (arts/entertainment and accommodation/food services), employment effects remain positive while sales effects attenuate and are less precisely estimated, suggesting that job growth is broader than the revenue gains concentrated in visitor-exposed sectors. Overall, the results provide rare establishment-level evidence that tribal casino development is associated with sustained private-sector expansion inside reservation boundaries, clarifying a key mechanism through which Indigenous self-determined enterprise can support long-run economic development.
Presented by: Joseph Aguilar-Bohorquez, University of Minnesota
The Effect of the EU CMRD on Firms’ Outcomes Abstract The Carcinogens, Mutagens or Reprotoxic Substances Directive (CMRD) sets binding occupational exposure limits (BOELVs) for 41 hazardous agents in the EU. While the directive is intended to protect workers’ health, its economic consequences for firms have received little attention. This paper exploits the staggered introduction of the 2017 and 2019 CMRD amendments across NACE-2-digit sectors to identify short-run firm-level effects using a modern Difference-in-Differences (DiD) estimator (Callaway & Sant’Anna, 2021). Treatments are defined by high-exposure divisions; controls are divisions not yet exposed. The analysis draws on Eurostat Structural Business Statistics (2005-2020) and maps each regulated substance to affected sectors. Results show temporary increases in (i) purchases of intermediate inputs, (ii) gross investment in tangible assets, and (iii) full-time-equivalent employment, with effects concentrated in the first two post-treatment years and fading thereafter. No statistically significant impacts are found on productivity, profitability, or the number of enterprises, indicating that compliance costs were absorbed through short-run input reallocation rather than persistent efficiency losses. Robustness checks (parallel-trend adjustments, Rambachan & Roth sensitivity analysis, exclusion of early-adopter Member States, and treatment-intensity specifications) confirm the direction of the main effects but reveal that causal inference relies on relatively strong parallel-trend assumptions. The study concludes that the CMRD induced modest, transitory adjustments at the division level without detectable medium-run deterioration in firm performance. These findings suggest that occupational-safety regulation can be compatible with maintaining competitiveness, provided firms have sufficient time to anticipate and implement compliance measures.
Presented by: Elena Meroni, European Commission, DG-JRC
Do Set-Aside Auctions Pay Off? Evidence from Brazil Abstract This paper investigates the impact of Set Aside (SA) auctions for Small and Medium Enterprises (SMEs) in Brazilian public procurement. Exploiting a legal reform that made SA auctions mandatory for contracts below BRL 80,000, we show that restricting competition to SMEs significantly increases their chances of winning contracts by 17.9\%, but also raises procurement costs by 12.1\%. This implies an annual fiscal cost of approximately BRL 73 million, given the total procurement spending affected by the policy. This result likely reflects the exclusion of larger, potentially more efficient firms. A trade-off emerges when analyzing the effects of mandatory SA auctions. Using a difference-in-differences approach, we find significant positive effects on employment and earnings for SMEs, with an average increase of 0.93 jobs per firm. Our cost-benefit analysis suggests that, despite the additional procurement expenses, the economic gains for SMEs indicate that SA auctions can be a viable policy tool in certain contexts. An estimated cost per job created of approximately USD 2,800.
Presented by: Alexsandros Cavgias Martins Frag, Ghent University
Session 15: Causal methods & Applications June 25, 2026 11:10 to 12:25 Location: Room 3
Session Chair: Martin Huber, University of Fribourg
Session type: contributed
Shapley Instruments Abstract When multiple valid instrumental variables are available, researchers must decide which instruments to include in their model to achieve optimal finite sample performance. We propose a novel instrument selection method based on the Shapley value. Our approach is examined under two different settings. In the first, there are many potentially weak instruments, and the goal is to select a subset that meets certain regularity conditions. In the second, we consider a heterogeneous treatment effect model and estimation of the Combined Compliers Local Average Treatment Effect (CC-LATE), where adding instruments entails a trade-off between improving precision and increasing the size of the Combined Compliers share. Preliminary simulation results demonstrate that our Shapley-based instrument selection method performs well in the many-weak-instruments setting.
Presented by: Giovanni Mellace, University of Southern Denmark
Identifying Treatment and Spillover Effects Using Control-Based and Forecast-Based Counterfactual Estimation Abstract Spillovers and interference pose fundamental challenges for causal inference, as treatment assigned to one unit may affect the outcome of others, violating the no-interference assumption underlying most empirical strategies. Existing approaches, based on partial interference, exposure mapping, spatial, network, or structural frameworks, typically rely on strong assumptions about interaction structures or require the existence of uncontaminated control units. We revisit this problem within the potential outcomes framework and compare identification under two broad classes of counterfactual methods: control-based counterfactual methods (CBCMs), such as difference-in-differences, and forecast-based counterfactual methods (FBCMs), including interrupted time-series and machine learning control methods. We show under which circumstances CBCMs and FBCMs identify average direct and spillover effects. Through simulations and an empirical application, we illustrate the advantages of FBCMs in the presence of pervasive spillover effects that fully invalidate CBCMs.
Presented by: Augusto Cerqua, Sapienza Università di Roma
Difference-in-Differences with Time-varying Continuous Treatments Using Double/Debiased Machine Learning Abstract We propose a difference-in-differences (DiD) framework designed for time-varying continuous treatments across multiple periods. Specifically, we estimate the average treatment effect on the treated (ATET) by comparing distinct non-zero treatment intensities. Identification rests on a conditional parallel trends assumption that accounts for observed covariates and past treatment histories. Our approach allows for lagged treatment effects and, in repeated cross-sectional settings, accommodates compositional changes in covariates. We develop kernel-based ATET estimators for both repeated cross-sections and panel data, leveraging the double/debiased machine learning framework to handle potentially high-dimensional covariates and histories. We establish the asymptotic properties of our estimators under mild regularity conditions and demonstrate via simulations that their undersmoothed versions perform well in finite samples. As an empirical illustration, we apply our estimator to assess the effect of the second-dose COVID-19 vaccination rate in Brazil and find that higher vaccination rates reduce COVID-19-related mortality after a lag of several weeks.
Presented by: Martin Huber, University of Fribourg
Session 16: Shocks & labour markets June 25, 2026 11:10 to 12:25 Location: Room 4
Session Chair: Thomas Baudin,
Session type: contributed
The Invisible Hand of the Caseworker: The Effectiveness of Support and Enforcement in Activation Abstract This paper investigates the causal effect of vocational training and sanction on employment outcomes using rich administrative data from Belgium. Our analysis tracks 48,249 individuals who entered unemployment between January 2022 and September 2023 over a period of up to 40 months. To address selection bias in training participation and sanction, we exploit quasi-random variation in caseworker assignment, leveraging differences in their preferences to refer jobseekers to training or sanction as an instrumental variable. Results show that vocational training delays re-entry into the labour market in the short term, particularly through a lock-in effect during the first year, but generates positive effects on stable employment from month 29 onwards. In contrast, sanctions lead to a short-term increase in temporary and regular employment and a higher likelihood of sector switching, but show limited long-term impact on employment. These findings provide new evidence on the comparative effectiveness of support-based and enforcement-based activation policies in fostering both timely labour market entry and sustained employment, while highlighting the crucial role of caseworker behaviour in shaping these outcomes.
Presented by: Ine De Mey, UGent
‘Dead Man Working’: A Place-based Approach to Occupational Safety and Health Abstract Despite increasingly stringent regulations, there has been a concerning stagnation in reducing workplace fatalities. Can place-based policy targeting help? By coupling machine learning techniques with comprehensive data from Italy, we develop a place-based approach to workplace fatalities. Harnessing accurate machine forecasts, we construct a granular risk map and compare it to the allocation of on-site work inspections and public subsidies for occupational safety, uncovering limited overlap. Counterfactual estimates reveal that current public policies are effective only in areas flagged as high-risk by ex-ante machine predictions. AI-powered territorial targeting can reduce the incidence of this chronic issue while lowering the costs of policy implementation. JEL codes:
Presented by: Marco Letta, Sapienza Università di Roma
Cultural Shock Transmission and Integration of Second Generation Immigrants Abstract Differences in norms and traditions between immigrants and their host society are known to shape inte- gration behaviours and the ways parents socialise their children. However, less is known about how the cultural distance between immigrants and natives at the time of arrival affects the integration of the second generation. Approaching culture through the lens of gender norms, this paper examines how the cultural shock experienced by first-generation migrants upon arrival shapes the multidimensional integration of their children in France. Our contribution is to measure cultural distance in a time-varying, arrival-cohort- specific way: using a dynamic index of gender norms, we construct the distance between France and each parent’s country of origin at the time of migration, allowing cultural exposure to differ across arrival cohorts within the same origin group. We show that greater parental cultural distance is associated with a weaker sense of belonging to France among second-generation immigrants, while having no detectable effect on labour-market outcomes. These relationships exhibit little heterogeneity by education or gender. Evidence on mechanisms points to the intergenerational transmission of more conservative views on gender roles and to larger attitudinal gaps with natives. Moreover, second-generation immigrants whose parents experienced a larger cultural shock appear to identify more strongly with their parents’ country of origin, consistent with a process of separation rather than marginalisation.
Presented by: Thomas Baudin,
Session 17: Migration June 25, 2026 14:45 to 16:00 Location: Room 1
Session Chair: Giacomo De Giorgi, IEE/GSEM University of Geneva
Session type: contributed
Signaling and Migrant Labor Market Integration: Experimental evidence from Colombia Abstract Labor market signaling plays a critical role in job matching by reducing information frictions for employers. This paper evaluates the effectiveness of signaling as a labor market integration policy by estimating heterogeneous returns to a skill certification program for both locals and migrants in Colombia. The program provides a common, credible signal to all participants, with access randomly allocated and 30 percent of spots reserved for migrants. Short-term, follow-up results indicate that certification increased employment among locals by reducing inactivity and increasing participation in informal work. For migrants, overall employment effects are negative but not statistically significant. However, formal employment increases substantially while skill downgrading significantly decreases. Additional evidence suggests that certification modestly raised migrants’ reservation wages, potentially contributing to short-term inactivity and unemployment.
Presented by: Juan Munoz Morales, IÉSEG School of management
Balancing the books: Immigration and local finances in France Abstract This paper studies the effect of immigration on local public finances in France, combining population census data with municipal budget accounts over the period 2007–2017. To address the endogeneity of immigrants’ location choices, I use an instrumental variables strategy based on historical settlement patterns by skill and nationality groups. The results show that immigration has a positive effect on per capita operating net revenues, as well as on both operating revenues and expenditures. Over the period under study, the immigrant share increased on average by 1.38 percentage points, from 8.63% to 10.00%, implying an estimated increase of 2.53% in total operating revenues per capita and 2.45% in total operating expenditures per capita. Immigration is also associated with higher investment activity. Overall, municipalities are able to maintain budgetary balance. Although migration is associated with higher debt levels, improvements in liquidity and self-financing capacity suggest that migration contributes to an expansion of local fiscal capacity.
Presented by: Nur Bilge, University of Lille, CNRS, IÉSEG School of Management, UMR 9221 Lille Economy Management (LEM)
Farming, Non-Farm Enterprises, and Migration under Incomplete Markets Abstract We study how rural households use non-farm entrepreneurship to smooth consumption following shocks to farm income. We show that transitory shocks to agricultural productivity lead farm households to reduce farm labor supply, increase non-farm labor supply, and become more likely to engage in non-farm entrepreneurship and temporary migration. Unlike temporary migration, these shocks have persistent effects on non-farm entrepreneurship, as engaging in non-farm activities enables households to accumulate skills that enhance their non-farm productivity. We then structurally estimate a dynamic model of household labor supply decisions across farming, non-farm activities, and temporary migration to produce policy relevant counterfactual exercises which reveal that over 30% of non-farm output comes is due to insurance motives. In terms of policy counterfactuals, we show that improving the functioning of insurance markets has substantial positive effects on structural transformation. Specifically, (index) weather insurance contracts and a minimum guaranteed income are welfare enhancing and allow people to engage in more productive activities. Furthermore, and of particular relevance, we study the potential implication of climate change, we find that a negative shift in the rainfall distribution leads to more than a 30% increase in non-farm output, and an increase in migration, underscoring the role of climate-driven structural transformation.
Presented by: Giacomo De Giorgi, IEE/GSEM University of Geneva
Session 18: Firms' finances June 25, 2026 14:45 to 16:00 Location: Room 2
Session Chair: Guglielmo Barone, Università di Bologna
Session type: contributed
Open Banking Regulation and Capital Allocation: Evidence from European PayTechs Abstract This paper investigates how regulatory reforms influence entrepreneurial finance and market dynamics in the financial technology (FinTech) sector. We study the impact of the Payment Services Directive 2 (PSD2), which introduced open banking across the European Economic Area (EEA), on venture capital (VC) investment in PayTech startups. Using VC funding data for PayTech firms in the EEA and the United States from 2010 to 2019, we apply a difference-in-differences approach to identify the causal effects of PSD2. We find that the Directive significantly increased both the volume and number of VC investments in EEA PayTechs, particularly among young, small, and B2B-oriented firms. The reform also stimulated early-stage funding and market entry, while its impact on exits remained limited. These results suggest that open banking regulation reduced compliance uncertainty and enhanced investor confidence, ultimately fostering entrepreneurial activity and capital allocation efficiency in the EEA FinTech ecosystem.
Presented by: Fernando Martínez-Santos, Joint Research Centre (European Commission)
The Value of Words: Evidence from Non-Financial Disclosure Regulation Abstract We examine the effects of laxer non-financial disclosure regulation on operating costs and access to external financing for micro firms in Italy. Starting in 2016, firms below certain size thresholds were exempted from filing reports with qualitative information that complements standard balance sheet items. Compliance rates were higher among older and more productive firms, in line with strategic considerations that play a role in policy adoption. However, the benefits of simplified reporting appear limited: in a regression discontinuity design that exploits the multidimensional size cut-offs that determine eligibility, we find no evidence of cost savings. We instead document a negative impact on ownership transfers and access to credit markets due to increased opacity, suggesting that the reduction of information disclosure to stakeholders may hinder business dynamism.
Presented by: Gianmarco Cariola, Banca d'Italia
Financial reporting accuracy and third-party auditing Abstract Credible financial accounts are essential for reducing information asymmetries in economic transactions. We introduce a novel measure of budget inaccuracy, defined as the difference between high-quality data on bank debt reported by the Credit Register and the corresponding liabilities recorded on balance sheets. Exploiting an audit reform in Italian municipalities and a difference-in-differences strategy, we show that randomly selecting auditors improves the quality of financial reporting. We also provide causal evidence that auditors’ characteristics, such as gender, experience, and social ties, which are randomly assigned to municipalities as a result of the reform, have no significant impact on audit quality.
Presented by: Guglielmo Barone, Università di Bologna
Session 19: Application of causal methods June 25, 2026 14:45 to 16:00 Location: Room 3
Session Chair: Jana Mareckova, University of St. Gallen
Session type: contributed
Reevaluating Causal Estimation Methods with Data from a Product Release Abstract Recent developments in causal machine learning methods have made it easier to estimate flexible relationships between confounders, treatments and outcomes, making unconfoundedness assumptions in causal analysis more palatable. How successful are these approaches in recovering ground truth baselines? In this paper we analyze a new data sample including an experimental rollout of a new feature at a large technology company and a simultaneous sample of users who endogenously opted into the feature. We find that recovering ground truth causal effects is feasible—but only with careful modeling choices. Our results build on the observational causal literature beginning with LaLonde (1986) offering best practices for more credible treatment effect estimation in modern, high-dimensional datasets.
Presented by: Justin Young, Microsoft
Group-Level Treatment Effect Heterogeneity in Difference-in-Differences: A Balanced Approach Abstract Understanding how treatment effects vary across groups is central to policy evaluation. In Difference-in-Differences designs, heterogeneity is often studied using subgroup or triple-difference analyses, which can suffer from conservative inference, reliance on parametric interaction structures, and sensitivity to differences in covariate distributions across groups. We propose the Balanced Group Average Treatment Effect on the Treated (BGATT), a new estimand that isolates heterogeneity in treatment responses from differences in covariate composition and is identified under standard conditional parallel-trends assumptions. BGATT provides a transparent target for comparing group-specific treatment effects. We derive an influence-function representation and develop estimators that are root-n consistent and asymptotically normal under flexible machine-learning estimation of high-dimensional nuisance components, enabling valid inference on both group-specific effects and differences across groups. Simulation evidence shows favorable finite-sample performance, and an application to the U.S. Medicaid expansion illustrates how BGATT clarifies whether observed differences across gender and age reflect true heterogeneity or compositional differences.
Presented by: Nadja van 't Hoff, University of Amsterdam
Fairness-Aware and Interpretable Policy Learning Abstract Fairness and interpretability play an important role in the adoption of decision-making algorithms across many application domains. These requirements are intended to avoid undesirable group differences and to alleviate concerns related to transparency. This paper proposes a framework that integrates fairness and interpretability into algorithmic decision making by combining data transformation with policy trees, a class of interpretable policy functions. First, dependencies between sensitive Attributes and decision-relevant features are removed through data pre-processing, then a Policy is obtained via tree-based optimization. Since data pre-processing compromises interpretability, an additional transformation maps the tree parameters back to the original feature space, yielding policy allocations that are pairwise independent of sensitive attributes and interpretable. The suggested approach is applied to the Problem of allocating unemployed individuals to different active labor market programs in Switzerland. It is demonstrated that, in this application, Fairness considerations can be integrated into an interpretable policy algorithm at comparatively low cost.
Presented by: Jana Mareckova, University of St. Gallen
Session 20: Intergenerational June 25, 2026 16:30 to 17:45 Location: Room 1
Session Chair: Paolo Ghinetti, Universita' degli Studi del Piemonte Orientale
Session type: contributed
Born Here, Raised There: Regional Mobility and Intergenerational Transmission of Education in Europe Abstract he effects of regional mobility during childhood on educational attainment and its transmission across generations remain understudied in Europe. Using micro-level data on three linked generations from the Survey of Health, Ageing and Retirement in Europe (SHARE), we track moves across European regions linked to historical GDP per capita at the regional level. For the parents' generation (G2), born between 1920–1950, moving to a better region before age 18 is associated with gains of up to 0.85 years of schooling. Conversely, for the children's generation (G3), born between 1960–1980, it is associated with a reduction of roughly 0.6 years of education, except when parents have tertiary education, in which case moving becomes beneficial. To achieve causality, we also adopt a sibling comparison strategy exploiting variation in the age of move and length of exposure which nets out all unobserved parental characteristics. We find that an additional year in a better region raises educational attainment by about 0.03 years. Overall we find evidence of a generational reversal indicating growing stratification by family background.
Presented by: Andrea Gasparroni, Università Politecnica delle Marche
University openings, college graduation, and labour outcomes in Italy Abstract This paper investigates the impact of tertiary education expansion on university graduation, labour market participation and employment in Italy. Exploiting differences in the timing of the first university faculty opening across provinces, we generate variation in age at exposure to higher education across birth cohorts within the same province. Linking cohorts born between 1950 and 1975 to local university availability, we estimate an event-study model to assess the long-run effects of increased access to tertiary education. We find that university expansion substantially increased graduation rates, with particularly strong effects for women. Labour market effects are more nuanced: exposure to local universities is associated with higher employment and lower inactivity at older ages, but gender differences in these outcomes are not always clearly distinguishable. Overall, our findings highlight the importance of geographic access to higher education for educational attainment and long-run labour market participation, especially in contexts with historically high female inactivity.
Presented by: Paolo Ghinetti, Universita' degli Studi del Piemonte Orientale
Session 21: Environment 1 June 25, 2026 16:30 to 17:45 Location: Room 3
Session Chair: Andrea Ciaccio, University of Padua
Session type: contributed
Induced Innovation and Collaboration: Insights from European Carbon Pricing Abstract This paper shows that the European Union Emissions Trading System (EU ETS) not only spurred low-carbon innovation among regulated firms but also reshaped R&D collaborations across firms and sectors. Using patent data matched with firm-level financials and ownership networks, I exploit the regulatory thresholds of the EU ETS to compare regulated and unregulated firms in a Difference-in-Differences design. ETS firms increase their number of low-carbon patents, with the largest gains during EU ETS phases II and III. Accounting for collaboration spillovers into the non-ETS group increases the estimated effect sizes. The policy also raises the share of patents developed in collaboration, particularly with non-competing firms that have high accumulated knowledge in low-carbon technologies. Collaborations with competing firms that were also subject to the carbon price or firms with lower knowledge remain largely unchanged. These results suggest that policies can generate innovation not only among directly regulated firms and sectors but also positive spillovers across the broader economy.
Presented by: Henri Gruhl,
Environmental Sustainability of Agro-Pastoral Development Interventions: Evidence from Tunisia Abstract This study evaluates the environmental sustainability of the Agro-pastoral Development and Local Initiatives Promotion Project II (PRODESUD II) in Tunisia using satellite-derived vegetation data and a Synthetic Difference-in-Differences (SDID) framework. Focusing on rangeland systems in semi-arid regions, the study constructs a composite Pasture Condition Score based on NDVI, EVI, and SAVI to capture multidimensional vegetation dynamics. Using grid-level panel data from 2008 to 2024, the analysis estimates both impacts during project implementation (2014–2020) and sustained effects in the post-completion period (2021–2024). Results indicate significant improvements in pasture conditions during implementation and a persistent, though attenuated, positive effect four years after project completion. The results provide empirical evidence that Agro-Pastoral Development interventions can generate durable ecological gains and demonstrate the value of integrating remote sensing data with quasi-experimental methods for ex-post sustainability assessment
Presented by: Ankhbayar Delgerchuluun, Center for Evaluation and Development
Difference-in-Differences under Neighbourhood Interference: An Application to the Health Effects of Regional Air Quality Policies in Lombardy Abstract We study identification, estimation, and inference for causal parameters in Difference-in-Differences (DiD) settings with multiple time periods, staggered treatment adoption, and spillover effects across units. Standard DiD approaches rely on the absence of interference, an assumption that is often violated when treatments propagate through economic, geographic, or social interactions. We consider a framework that allows for spillovers without requiring knowledge of the underlying interference structure. We propose a class of causal parameters that separately characterize direct treatment effects, spillover effects, and treatment effects under spillover exposure, and show that they are nonparametrically identified under conditional parallel trends assumptions accounting for both treatment timing and spillover exposure. Our identification strategy exploits the presence of units that are not exposed to spillovers to recover counterfactual outcomes for exposed units. The identification results lead to a family of estimators based on a double inverse probability weighting scheme that reweights observations according to treatment timing and spillover exposure probabilities. We establish consistency and asymptotic normality of the proposed estimators under approximate neighbourhood interference, allowing for weak dependence across units in a network, and propose a multiplier bootstrap procedure for asymptotically valid simultaneous inference. Finite-sample performance is examined through Monte Carlo simulations, and the proposed methodology is illustrated in an empirical application evaluating the impact of traffic restrictions introduced under a regional air quality plan on health outcomes using environmental and health data from Lombardy, Italy.
Presented by: Andrea Ciaccio, University of Padua
Session 22: Higher Education 2 June 25, 2026 16:30 to 17:45 Location: Room 2
Session Chair: Elena Lazzaro,
Session type: contributed
The Timing of Academic Specialization: Impacts on Student Outcomes Abstract Giving students an opportunity to explore different fields before committing to a discipline in college is likely to improve match quality between students and their chosen fields of study. However, students who specialize earlier can pursue a more coherent and focused curriculum, better aligning their academic record with future educational or occupational aspirations. This paper presents the first causal evidence on the effects of this trade-off in the timing of specialization in a setting where students have almost completely free course choice. We use an instrumental variable approach, exploiting course prerequisite structures and timetable variation, to evaluate impacts on curriculum composition, college performance, post-graduate degrees, and labor market outcomes. We find that students who specialize early have a more focused curriculum and a higher likelihood of pursuing relevant post-college degrees, but they do not outperform late-specialized peers either during or after college.
Presented by: Tina Rozsos, Vrije Universiteit Amsterdam
Higher Education Expansion and the Upward Mobility of Immigrant Families in France Abstract This paper examines the democratization effect of higher education expansion with a focus on immigrants’ children. I combine a newly constructed dataset on higher education institutions in France with individual-level administrative data. Applying a matching strategy and a two-way-fixed-effects design interacted with a parental-origin indicator, I assess the differential causal impact of higher education departments creations in the 1990’s on a new individual-level indicator of intergenerational social mobility, by parental origin. Children of immigrants see a 2.8 percentile greater increase in the income percentile difference with their parents than children of native-born parents. Further evidence indicates that the expansion was particularly advantageous for children of foreign-born parents located around the middle of the income distribution. The underlying mechanisms include local externalities, the creation of higher education units in more isolated territories, as well as increased access to STEM education. While the results vary according to the country of birth of the parents, no significant differences are observed by gender.
Presented by: Mikaël Pasternak, University of Lille
Excellence Funds, Excellent Outcomes? Lessons from Italian University Departments Abstract This paper investigates the causal effects of a competitive excellence initiative financing Italy’s top-25% university departments. We exploit a novel linkage of administrative data on academics’ careers, publications, departmental activities, and student enrollment to estimate both the direct and indirect effects of the policy in a dynamic difference-in-differences framework. The additional financial resources directly allowed funded departments to increase their faculty size, but had a limited impact on educational offerings. Indirectly, the excellence designation led to a higher enrollment of students, particularly high-achieving ones, suggesting a positive quality signal. The funding boosted scientific output but not productivity, except in large STEM and Life Sciences departments, which benefited from agglomeration effects and improved research infrastructure.
Presented by: Elena Lazzaro,
Session 23: Wages June 26, 2026 9:00 to 10:40 Location: Room 1
Session Chair: Massimiliano Bratti, University of Milan
Session type: contributed
Working hours policy reform: micro and macro impacts Abstract We study the macroeconomic consequences of reducing legal working hours, using Chile’s 2005 reform as a quasi-natural experiment. While traditional microeconometric evaluations suggest nonsignificant impacts on directly affected workers, we argue that these comparisons fail to capture the full policy effect because they overlook the "missing intercept"—the aggregate shift in market conditions that affects both treated and control groups simultaneously. Since such reforms influence nearly 60\% of the workforce, they trigger general equilibrium spillovers that fundamentally alter the outside options and bargaining power of all individuals. Using instrumental variable local projections, we find that the reduction in hours increased total employment by 1.9 percentage points, primarily driven by a 1.4 percentage point surge in self-employment as salaried private positions decrease, as they have become more costly. These results, rationalized through a structural search and matching model, show that legal hour regulations reshape labor market composition, cautioning policymakers against relying on micro-level evidence, which may provide a misleadingly narrow picture of aggregate impacts.
Presented by: Mauricio Tejada, Universidad Diego Portales
The Earnings Premium for Long Hours: A Directed Search Approach Abstract There are substantial differences across occupations in the earnings premium for working long weekly hours, causing gender-biased sorting and inequality in the labor market. To explain why these premiums emerge, previous work emphasizes the role of firm-side production technologies. However, hours and wages are equilibrium outcomes, which also depend on worker-side preferences and the labor market structure. To study the importance of these forces, we build a tractable directed search and matching framework with two-sided multidimensional heterogeneity, where wages and hours are jointly endogenous. We estimate the model using data representative for the upper end United States labor market. We find that productivity differences are important in shaping the large earnings premiums in Business and Law occupations. Yet to a large extent, the earnings premium simply compensates workers for the additional disutility associated with long working hours. In tighter markets, such as Tech, the added bargaining power of workers inflates the premium. Counterfactual simulations show that removing firms’ incentives to reward long hours reduces the gender gap in earnings marginally, while equalizing the worker-side costs of working long hours almost halves the gap.
Presented by: Tom Potoms, University of Sussex
Firm Recruitment under Guaranteed Minimum Income: Evidence from Italy’s Online Job Vacancy Data Abstract This paper studies the effects of Guaranteed Minimum Income (GMI) schemes on firm recruitment, focusing on the Italian case. Leveraging rich Online Job Vacancies (OJV) data, we examine how the introduction of the GMI affects the number, duration, and characteristics of job postings. We imple- ment a Difference-in-IV strategy, exploiting municipal variation in GMI exposure after its introduction. This approach allows us to identify the causal impact of GMI on formal firm recruitment behavior. We find that firms more heavily exposed to the GMI experience an increase in the number and duration of postings, as well as in the likelihood of disclosing wages. We also find that the effects of the GMI are highly heterogeneous depending on the firm- and province-specific characteristics. Our findings contribute to a better understanding of how GMI schemes may affect firm recruitment dynamics.
Presented by: Massimiliano Bratti, University of Milan
Session 24: Labour 2 June 26, 2026 9:00 to 10:40 Location: Room 2
Session Chair: Michele Pellizzari, University of Geneva
Session type: contributed
Overpromising Green Jobs? Evidence from the French Energy Efficiency Obligations Program Abstract The EU Green Deal promises one million jobs by 2030, with energy efficiency as key driver. However, projections rely on unverified ex-ante estimates. This paper provides the first ex-post estimate of employment impacts from a large-scale energy efficiency programme. Using a policy discontinuity in France and a state-of-the-art synthetic control method on disaggregated data, we find 2.2 job-years created per million euro invested, far below the 8.52 jobs assumed in EU assessments. We find no evidence of a major labor shortage, as transition costs are low and wages remain stable. However, only 10% of the subsidy increase translates into new hires, while 28% is absorbed into firms’ value added. Hence, rent capture by local energy-efficiency installers leads to an inefficiency in the use of the subsidies, potentially undermining the green transition.
Presented by: Guillaume Wald, Mines Paris - PSL
Early Arrivals: How Preterm Birth Affects Mothers and the Role of Firms Abstract This paper studies the labor market consequences of preterm birth for mothers and fathers, and investigates the role of workplace stress, proxied by firm-level mass layoffs, in causing adverse birth outcomes. Using administrative data from Italian social security records, we identify the causal effect of preterm birth through two complementary strategies: a matched event-study design that pairs each preterm birth to an observationally equivalent term birth on rich pre-conception maternal, paternal, and firm characteristics, and a within-parent design that exploits variation in gestational outcomes across a mother’s own pregnancies. We find that preterm births at or below 32 weeks of gestation reduce maternal log earnings by approximately 17 log points at the second post-birth year, with only partial recovery thereafter; for the most severe cases (≤ 28 weeks) the penalty at the same horizon roughly doubles. Fathers bear negligible costs. The penalty is mediated through reductions in weeks worked and exit from employment, and is amplified among mothers whose partners have low earnings and among those employed in high-pressure workplaces. Turning to causes, we exploit firm-level mass layoffs as a source of plausibly exogenous variation in workplace stress. Because Italian law prohibits the dismissal of pregnant workers even during collective redundancy procedures, any effect on birth outcomes must operate through a non-pecuniary stress pathway rather than own income loss. Consistent with a financial-anxiety mechanism, the effect is entirely concentrated among mothers in below-median-income households and is statistically indistinguishable from zero elsewhere. The two parts of the analy- sis together reveal a compounding inequality mechanism: socioeconomic disadvantage raises the probability of severe preterm birth, which in turn deepens the maternal earnings penalty, with both margins falling disproportionately on the same population.
Presented by: Joanna Kopinska, Sapienza Università di Roma
The Impact of Digital Financial Transactions on Women’s Bargaining Power: Evidence from India Abstract After introducing the Digital India programme, India experienced rapid mobile penetration and growth in internet services. Despite having access to digital devices and various payment apps, women still face difficulties negotiating with their husbands and other family members. This paper examines the impact of digital financial transactions (DFT) using mobile phones on women’s bargaining power in Indian households, based on the National Family Health Survey-5 (2019-21). We applied a conditional mixed-process model to correct for selective bias and unobserved heterogeneity. We employed the instrumental variable approach, using a leave-one-out rate, to estimate the causal effect of DFT on women’s household bargaining power. Our findings indicate that married women using mobile money have a positive, statistically significant impact on their bargaining power. The subsample analysis reveals that the impacts vary among families with distinct characteristics, including wealth status, caste, and head of household. Our study is robust, incorporating village-level instrumental variables and age groups, and divides the sample into distinct subgroups to highlight the role of women in household decision-making in the digital age, aligning with Sustainable Development Goal 5 for gender equality.
Presented by: Shruti Sengupta, IIT Roorkee
Protecting domestic workers from migration: evidence from an unconventional Swiss policy Abstract We study the labour market effects of a migration-related policy introduced in Switzerland in 2018 to protect domestic jobseekers. The policy requires firms with vacancies in selected occupations to notify these vacancies to the public employment service and prohibits advertising them elsewhere for five days. During this period, registered jobseekers have exclusive access to the vacancies. Using administrative data covering the universe of registered jobseekers, we find that referrals to protected vacancies increase job-finding probabilities, particularly when re-employment occurs in the same occupation. However, because targeted occupations are characterised by high separation rates, these short-run gains are offset by higher subsequent unemployment persistence. We also document that the policy substantially increases workload for local public employment offices, leading to a reallocation of caseworker effort toward protected occupations and generating negative spillover effects on jobseekers from non-targeted occupations.
Presented by: Michele Pellizzari, University of Geneva
Session 25: Shocks & mobility June 26, 2026 9:00 to 10:40 Location: Room 3
Session Chair: Fabio Berton, European Commission, Joint Research Centre
Session type: contributed
Migration Policy Backlash, Identity and Integration of Second-Generation Migrants in France Abstract Do symbolic aspects of integration policies affect migrants’ integration into the host society? In this paper, we study the effects of a symbolic change in birthright citizenship rules in France that requires second-generation immigrants to state their allegiance on their integration. Adopting a Difference-in-Differences approach, we show that, contrary to its stated aim of fostering a greater sense of belonging, this policy led to a loss of national identity and an increase in perceptions of discrimination among the target group. We document that these effects are not driven by changes in naturalization rates or an increased general hostility. We also show that while the reform did not affect their economic or political integration, it did reduce their cultural integration, as measured by religiosity and naming patterns. Overall, rather than promoting integration, such migration policies can lead to a backlash.
Presented by: Simone Moriconi, IESEG School of Management
I'll be there for you: natural disasters, aid, and trust in public institutions Abstract This paper examines whether receiving financial assistance from the European Union (EU) after natural disasters during early adulthood shapes individuals' subsequent confidence in the EU. We focus on the European Solidarity Fund (EUSF) and combine individual-level survey data with information on disasters and EUSF interventions over the 2002–2018 period. Exploiting the conditional exogeneity of natural disasters and geocoded data on their subnational scope, we show that EUSF aid increases citizens’ confidence in the EU. The effect is stronger in regions with lower-quality local governance and among more risk-averse individuals. Finally, we show that the EUSF mitigates the economic downturns that typically follow disasters, providing evidence of both political and economic returns to EU-level solidarity.
Presented by: Giulia Romani, Joint Research Centre
Do Commuting Subsidies Drive Workers to Better Firms? Abstract Commuting subsidies may improve job match quality by expanding the choice set of feasible job opportunities for workers. Using variation in tax rates, initial commuting distances, and major reforms to Germany’s commuting subsidy formula, we construct worker-specific changes in subsidy generosity. More generous subsidies induce workers to commute longer and switch to higher-paying firms, implying a positive causal effect of commuting distance on the firm-specific component of pay. Commuting subsidies also strengthen positive assortative matching by reallocating high-ability workers to more productive plants. While subsidies raise wages on average, the resulting increase in assortativity contributes to higher earnings inequality.
Presented by: Elke Jahn, IAB, Bayreuth University
Strangers in the night. Should we really be scared by migrant workers? Abstract The aim of this paper is to assess the causal impact of foreign labor supply shocks to net and gross employment changes and on workforce composition. To do so, we take advantage of the rules governing legal workers’ flows in Italy, where each year the Ministry of Domestic Affairs sets an overall quota of migrants awarded to applicant employers based on a first-come, first-served rule. By propensity-score-matching observed successful applicants are matched to the other potential employers and using an unprecedented combination of different administrative data sources, we find that the arrival of migrants enhances net employment flows by one low-educated, old-aged domestic worker every six firms. Moreover, we find that the null net effects observed on the other workers’ breakdowns follow from significant effects on gross in- and out flows that cancel out each other, thus highlighting a fundamental role of migrant workers’ supply shocks as trigger for workers’ turnover.
Presented by: Fabio Berton, European Commission, Joint Research Centre
Session 26: Public good June 26, 2026 9:00 to 10:40 Location: Room 4
Session Chair: Enkelejda Havari, IESEG School of Management
Session type: contributed
Lobbying Strategy in the Age of Artificial Intelligence Abstract This study examines how rising government attention to artificial intelligence (AI) reshapes corporate political mobilization. As AI becomes increasingly central to public policy debates, firms exposed to AI-related technological change face stronger incentives to monitor and influence emerging regulatory agendas. Using firm-level lobbying data and a measure of government AI attention based on U.S. congressional discourse, the analysis investigates whether AI-exposed firms adjust their lobbying behavior when AI gains political salience. Preliminary results show that increases in government AI attention are followed by stronger lobbying responses among medium- and high-AI firms, while contemporaneous effects are weaker. Evidence around the 2019 American AI Initiative further suggests that AI-exposed firms intensified political mobilization after AI became a more explicit national policy priority. However, these lobbying responses do not translate into clear short-term financial performance gains. The study therefore suggests that AI-related lobbying may matter less through immediate profitability than through longer-term advantages in regulatory access, market positioning, and the reinforcing relationship between political power, economic power, and market dominance.
Presented by: Wenjian WANG, Skema Business School; Université Côte d'Azur
Urban Water Rationing and Domestic Violence: Evidence from a Spatial Regression Discontinuity Design Abstract Does resource scarcity inevitably trigger intra-household conflict? While a large literature documents links between rainfall shocks and violence in rural economies, evidence on the social consequences of infrastructure-driven scarcity in urban settings remains limited. This paper examines the causal effect of urban water rationing on domestic violence using a large-scale natural experiment from Bogotá, Colombia. In response to the 2023–2024 El Niño episode, the city implemented a deterministic system of 24-hour water service suspensions that rotated across nine fixed hydraulic service zones, generating sharp spatial and temporal discontinuities in access to water. I exploit this variation using a boundary-specific Spatial Regression Discontinuity Design that compares reported domestic violence incidents in treated and untreated neighborhoods located immediately across hydraulic service boundaries on the same day. The analysis combines high-frequency, georeferenced administrative data from Bogotá’s emergency response system with detailed information on rationing schedules and network topology. Estimation proceeds by recovering local treatment effects at each spatial cutoff and aggregating them using precision-weighted methods, allowing for heterogeneity and valid inference under spatial dependence. Across more than 20,000 incidents and 16 treatment–control boundaries, I find no evidence that temporary water rationing increases reported domestic violence. Boundary-specific estimates are tightly centered around zero, and the pooled effect is small and statistically indistinguishable from zero. The results are robust to alternative bandwidth choices, donut regression discontinuity specifications, and placebo tests at artificial cutoffs, and are sufficiently precise to reject meaningful increases in violence. These findings contrast with evidence from rural settings, where water scarcity often operates through income shocks and has been linked to heightened intimate partner violence. The results suggest that short, predictable interruptions to urban water service do not generate comparable social spillovers. Methodologically, the paper contributes a flexible spatial regression discontinuity framework applicable to urban policies implemented through fixed infrastructure networks. Substantively, it provides new evidence on the social impacts of demand-side water management in cities facing climate-related supply stress.
Presented by: Natalia Galvis Arias, University of Manchester
Do Incentives and Organizational Innovations Affect the Performance of Public Administrations? Lessons from the Italian NRRP Abstract We study how public institutions respond to incentives and organizational innovations in the provision of infrastructure, a core public good. We focus on Italy’s National Recovery and Resilience Plan (NRRP), which combines an unprecedented investment program with a distinctive governance model for project delivery. NRRP projects were subject to: (i) performance-based conditionality tied to the achievement of milestones and targets; (ii) intensive internal and external monitoring along predefined timelines; and (iii) a dedicated regulatory framework aimed at streamlining procedures and strengthening the administrative capacity of procuring entities. Exploiting variation within and across municipalities, as well as revisions to the Plan that altered projects’ treatment status over time, we identify the effects of each channel. We find that the NRRP significantly accelerated both the awarding and the execution of public works. In the awarding phase, performance-oriented approach played a central role in improving administrative performance. In the execution phase, the timely availability of funding was the key driver of the rapid start of construction, reflecting that success of the project delivery depends both on public and private effort. Moreover, we detect positive spillovers on the standard procurement activity of NRRP-exposed municipalities, particularly among administrations with lower initial institutional capacity, suggesting a steep learning process induced by the Plan’s organizational innovations.
Presented by: Antonio Accetturo, Banca d'Italia
Faith, Interrupted: Identity and Behavior After Forced Atheism Abstract Governments have long intervened in national identity formation, yet it remains unclear whether such efforts truly reshape values or primarily induce compliance. This paper studies Albania’s 1967 Cultural Revolution, when the Communist regime criminalized religion nationwide and promoted a secular “pure Albanian” identity. We exploit this reform as a nationwide shock to religiosity and examine its effects on public and private expressions of identity, values, and socioeconomic behavior. Combining administrative and survey data and applying a residualized event-study design, we show that religious names declined by 50–150%, while secular Albanian names increased by 50%. These shifts were largest in more illiterate districts, consistent with stronger exposure to state control, yet substantial resistance persisted: religious minorities were more likely to continue religious naming practices. Survey evidence reveals large declines in religious participation and upbringing among those born after the reform, but no change in belief in God or broader political values. This pattern suggests that the reform affected observable behavior and identity signaling, but no deeper ideological shifts.
Presented by: Enkelejda Havari, IESEG School of Management
Session 27: Labour & shocks June 26, 2026 11:10 to 12:25 Location: Room 1
Session Chair: Marco Visentin, University of Zurich
Session type: contributed
Understanding Workers’ Migration Response to Shocks: Mass Layoffs in Brazil Abstract This paper investigates the impact of unexpected employment shocks on workers' migration decisions using matched employer-employee data from Brazil's formal sector. Leveraging an event-study design, we find that workers who separate from their firm during a mass layoff face a significant and persistent 10% decline in their earnings relative to non-laid-off workers. We also find that laid-off workers are 15 percentage points more likely to migrate to another labor market. While all laid-off workers migrate at higher rates than non-laid-off workers, we find that men migrate at higher rates than women, and that non-white workers migrate more than white workers. We explain these patterns through a descriptive exercise comparing wage trajectories of migrant and non-migrant laid-off workers. We find 1.5% higher earnings among migrants than stayers, with almost all gains concentrated among men and larger gains among non-white workers. We complement our empirical evidence with a location-choice model that delivers a gravity equation for bilateral migration flows. Gravity estimates indicate that a one-standard-deviation increase in layoff intensity raises flows out of the origin by roughly 1-3%. Taken together, the findings show that mass layoffs trigger substantial but selective geographic reallocation, and that unequal access to migration shapes the distributional consequences of employment shocks.
Presented by: Takshil Sachdev, Cornell University
Climate Stress and International Migration from MENA Abstract Existing research on the drivers of international migration is constrained by incomplete measurement of volumes, directions and country coverage. Most studies focus on recent decades, a narrow set of OECD destinations or a single flow type, even though categories are blurred, new destinations have emerged and many developing countries now host large migrant populations. Recent work suggests that migration drivers are best understood within a broader migration system rather than in isolation. However, even in this newer literature, long-run historical ties, such as deep ancestry links or very old migrant stocks, are still rarely analyzed jointly with contemporary drivers and environmental pressures in a unified empirical framework. In this paper, we focus on MENA countries and use bilateral flow estimates from Abel (2018) in a gravity framework to analyse how environmental stress interacts with the broader social, political, and historical structures that organise migration. We distinguish fast-onset shocks, measured by the number of people affected by floods and storms, from slow-onset anomalies, captured by 12-month SPEI indicators for unusually dry and unusually wet con- ditions. These origin-year shocks have been analyzed in a rich context of civil conflict, development, urbanization, and agrarian exposure, and we further augmented the study with historical migrant stocks around 1960 and a new ancestry stock measure we devel- oped, reflects historical ties built up in the Colonial Era. This study is one of the first to incorporate deep ancestry-based ties into an empirical analysis of migration from MENA. Our results show that fast-onset disasters affect emigration from MENA in a non-linear way: small and moderate events are largely absorbed locally, while very severe shocks lead to sharp increases in cross-border outflows. Deviations from normal moisture in either direction, whether drought or excess rainfall, are also associated with higher emigration. In addition, migrant stocks and ancestry links are strong and robust predictors of where migrants move, even after we control for other conditions.
Presented by: Mehmet Burak Akren, Duisburg-Essen University
Subcontracting, Procurement Performance & Workplace Safety Abstract Subcontracting in public procurement may alleviate execution frictions, by expanding the firm boundaries or by insuring it against unforeseen circumstances that obstacle timely and complete project delivery. However, the delegation of tasks to external firms may dilute legal responsibility, generate uncertainty over ultimate accountability and weaken incentives to safeguard workplace safety. To investigate this trade-off, we draw on comprehensive data covering the universe of public works tenders in Italy, enriched with information on contractors, subcontractors, and workplace accidents. Leveraging a difference-in-differences design that exploits a 2021 reform easing restrictions on subcontracting, we show that increased access to subcontracting improved project execution by reducing public works' completion time and cost overruns, despite the apparently lower competition at the tender stage. On the downside, within-firm estimates indicate that the incidence of work-related accidents may be higher when a firm operates as a subcontractor, illustrating a policy-relevant tension between effective execution and work-place safety.
Presented by: Marco Visentin, University of Zurich
Session 28: ALMP June 26, 2026 11:10 to 12:25 Location: Room 2
Session Chair: Bart Cockx, Ghent University
Session type: contributed
Supporting Mothers Back to Work: Experimental Evidence on Employment, Fertility, and Child Outcomes Abstract Many advanced economies face persistently low fertility alongside rapid population ageing, raising concerns about economic sustainability and demographic balance. Addressing these challenges requires both sustained labor market participation among the working-age population and conditions that support childbearing. These objectives place women, and particularly mothers, at the center of the demographic debate, as motherhood remains a key turning point in employment trajectories and family formation. Using experimental evidence from an intervention targeting mothers who curtailed employment due to childcare responsibilities, the paper finds that improving work–family reconciliation can support mothers’ labor market reintegration, promote investments in existing children, and, under conditions of greater stability, strengthen fertility desires.
Presented by: Chiara Pronzato, University of Turin and CCA
Cash Transfers and Parental Investments: Evidence from the Bolsa Familia Program Abstract This paper investigates whether Conditional Cash Transfer (CCT) programs can impact parental involvement in children’s education, a key factor in human capital accumulation that remains understudied in this policy context. I study this question considering Brazil’s Bolsa Familia Program, the world’s largest CCT. Exploiting an exogenous expansion in program coverage and a methodological change in the allocation rule of new slots to municipalities, I estimate the impacts of increased exposure to the program using a difference-in-differences approach. Drawing on administrative education data from the Sao Paulo state, I find that the Bolsa Familia expansion enhanced parental supervision of children’s school-related activities and increased students’ participation in the State exam. These behavioural changes seem to be associated with measurable improvements in students’ performance, notably a higher likelihood of passing the 9th grade and an increase of approximately 4% of a standard deviation in standardised scores in Portuguese.
Presented by: Daniele Chiavenato, Nova School of Business and Economics
Biased Beliefs in Job Search: Structural Evaluation of Coaching and Counseling Abstract Using Swiss data on job search behavior and expectations about interviews, we evaluate a program combining intensified counseling and coaching to address biased beliefs, in a randomized trial. We develop a job search model with Bayesian updating of biased perceptions about the number of interviews. Job seekers are initially overoptimistic, then become over-pessimistic as past signals fade. The model replicates the non-monotonic job-finding rate and patterns in applications, (expected) interviews, and wages. It decomposes the treatment effect into key drivers---search efficiency, job search costs, and beliefs---identifying belief formation as main channel through which the program accelerates employment transitions.
Presented by: Bart Cockx, Ghent University
Session 29: Environment 2 June 26, 2026 11:10 to 12:25 Location: Room 3
Session Chair: Valentin Cocco,
Session type: contributed
Assessing the Impact of Low-Emission Zones on Urban Air Quality: Evidence from Madrid Abstract This paper analyses the impact of Low Emission Zones (LEZ) in Madrid, focusing on two sequential policies: Madrid Central and Madrid 360. Extending prior studies that assess short-term effects of Madrid Central up to 2019, this research contributes by incorporating updated panel data until 2024, thereby encompassing the full implementation of Madrid 360 and the COVID-19 period. Using a Difference-in-Differences (DiD) strategy on data from 21 air quality monitoring stations across Madrid, we estimate the causal effect of these interventions on pollutant concentrations, particularly NO2, PM2.5 and PM10. The results provide insights into the long-term effectiveness of LEZs and the importance of policy continuity and design in achieving European air quality standards.
Presented by: Rocío Sánchez-Mangas, Universidad Autónoma de Madrid
Manipulation in rural landscapes: farmers’ strategic response to environmental policies Abstract To promote more sustainable agriculture, the European Union's Common Agricultural Policy tied 30% of farmers’ subsidies to allocating 5% of arable land to biodiversity-friendly landscape features like fallows, grass strips, hedgerows, or catch crops. However, farmers could exempt from the requirement depending on their position relative to three thresholds. Using plot-level land-use data of all French farms applying to CAP subsidies, we show that farmers endogenized these exemption rules, as we detect manipulation of the assignment variables for 20% of the just-exempted farms. Because such strategic sorting invalidates standard Regression Discontinuity Designs (RDD), we apply manipulation-robust RDD at the thresholds to estimate the reform’s causal effects on farms’ EFA cover. Our findings reveal limited additionality: any apparent increases in EFA cannot be distinguished between genuine improvements and farmers’ strategic responses to exemption rules. The only robust increase concern farms that were below but close to meeting the 5% requirement. Moreover, this increase appears to be supported exclusively by catch crops, suggesting that compliance was concentrated on the least ecologically valuable option.
Presented by: Valentin Cocco,
Supply Chain Linkages and Carbon Leakage under the EU ETS Abstract This paper examines how carbon pricing propagates through global production networks and conditions carbon leakage in trade. Using sector-level bilateral intermediate trade data from 2000–2022, we estimate a structural gravity model with exporter–sector–year, importer–sector–year, and bilateral fixed effects. Identification exploits the staggered introduction and sectoral coverage of the EU Emissions Trading System (ETS).Baseline estimates show that imports from non-ETS exporters to ETS-regulated importer–sectors decline by about 24 percent, indicating no average trade-based leakage. However, allowing for heterogeneous effects by input–output exposure reveals strong network dependence. The leakage effect decreases with both upstream and downstream supply-chain integration and becomes negligible or negative in highly network-embedded sectors. These findings suggest that carbon pricing shocks transmit along global value chains, attenuating leakage in deeply integrated industries. The results highlight the importance of production networks for understanding the international incidence of climate policy.
Presented by: Lewoye Bantie Baylie, Università di Salerno

 

This program was last updated on 2026-07-03 11:59:32 EDT