Peer-Reviewed Publications

1. Pisanelli, Elena. “A Tale of Two: Can Gender Equality and Efficiency Go Hand in Hand?” Italian Economic Journal, 2025.

This paper examines the tension between efficiency and gender equality in hiring. Companies aim to streamline processes while ensuring fairness, particularly in achieving gender balance. Despite advancements in hiring algorithms, studies suggest lingering biases against women. Using real-world recruitment data, I analyze the gender equity implications of two popular hiring tools: predictive algorithms and assessment software in full automation. I replicate two widely used resume screening mechanisms—one predictive, the other assessment-based—and contrast them with human recruiters. The findings reveal a significant discrepancy: predictive algorithms maintain gender inequality and mimic human recruiters’ selections, while assessment software significantly boosts female candidates' chances by 79 percentage points (with a baseline selection rate of 22 percentage points for female applicants) and enhances the overall qualifications of the selected pool. These results challenge existing research on AI-driven hiring, suggesting a cautiously optimistic approach to exploring hiring algorithms’ role in fostering gender equality.

2. Pisanelli, Elena. “Divorce, domestic violence, and help-seeking” Italian Economic Journal, 2024.

This paper examines the impact of the 2014 Italian divorce law on help-seeking behavior of domestic violence victims and femicides. The paper finds that contrary to expectations, the reform, which aimed to reduce the cost of divorce while requiring mutual consent, led to a decrease in help-seeking behavior among intimate partner violence survivors and an increase in femicides perpetrated by husbands. These findings suggest that while reducing the cost of divorce may empower individuals to leave abusive relationships, the requirement for mutual consent may inadvertently escalate violence as husbands seek to assert control and prevent separation. The study underscores the importance of considering the unintended consequences of divorce legislation and prioritizing the safety of intimate partner violence survivors in family policy interventions.

3. Pisanelli, Elena. “Your resume is your gatekeeper: Automated resume screening as a strategy to reduce gender gaps in hiring” Economics Letters, 2022, vol. 221, 110892. 

Firms increasingly rely on artificial intelligence (AI) algorithms for hiring. The literature prompts concerns about such AI algorithms hindering gender equality in employment outcomes. Using a unique field study, I find human recruiters’ gender stereotypes lead to women having 69% lower chances of being interviewed for a gender-neutral job, compared to equally qualified men. Introducing automated resume screening shrinks such a gender gap by 43 percentage points.

Submitted Working Papers

1. Pisanelli, E. "Coercion, Reporting, and Colonial Rule: Intimate Partner Violence across History Evidence from India" 

This paper estimates the long-run legacy of colonial governance for coercion within households. I exploit a sharp historical boundary in South India that placed otherwise neighboring districts under two distinct institutional regimes: direct British administration in the Madras Presidency and indirect rule in the Princely State of Hyderabad. Using a geographic regression discontinuity design with georeferenced village clusters from India’s Demographic and Health Surveys, I document a discrete increase in intimate partner violence on the former British side of the border. The margin of violence also changes: physical and sexual abuse rise, while threats decline. To probe mechanisms and measurement, I show that women exposed to the former British regime are less likely to seek help and more likely to view wife-beating as justified, consistent with a persistent gap between incidence and disclosure. I then triangulate household survey measures with administrative crime statistics. Results show lower reporting or different engagement with state institutions despite higher survey-measured violence. The findings show that colonial-era institutional environments can persistently shape both the prevalence of interpersonal violence and the visibility of that violence in official records, with implications for how economists interpret administrative data on crime and household welfare.

2. Pisanelli, E. "From Headlines to Home: Femicide News, Paternity Leave, and Men's Responses to Intimate Partner Violence", Available at SSRN: https://ssrn.com/abstract=6148946

This paper studies how Italian fathers respond to intimate-partner violence (IPV) when severe violence becomes salient in their information environment and when family policy reshapes fathers’ roles at home. I combine an original survey with (i) a vignette experiment that varies the form of IPV (physical assault versus psychological control) and randomly provides information on IPV prevalence and support services; (ii) a municipality-by-day panel of local femicide news (Corriere della Sera Pro, 2022–2024); and (iii) a regression discontinuity design around the January 1, 2021 expansion of mandatory paid paternity leave from one to ten days. Three findings emerge. First, fathers systematically discount non-physical IPV: psychological control is perceived as less serious and elicits more victim-blaming and less perpetrator-blaming. Information provision modestly improves selected pro-victim and institutional-trust outcomes but does not close this gap. Second, cumulative femicide exposure is only weakly related to attitudes, but interviews conducted immediately after a local femicide show large, short-lived shifts in perceived attention to male victims and in selected masculinity norms, with effects that differ by vignette type and information-treatment status. Third, eligibility for expanded paternity leave is associated with more egalitarian parenting norms and a marked reduction in the perception that male victims are neglected; these impacts are likewise heterogeneous across the experimental arms. Overall, the results distinguish transient salience responses from slower, role-based policy effects and speak to the design of publicly funded IPV prevention, support services, and family policy. Effects are generally larger among fathers with above-median pre-treatment emotional load (and, in some cases, above-median within-couple emotional-load gaps), consistent with role-conflict and stress mechanisms.

3. Barigozzi, F, P Biroli, C Monfardini, N Montinari, E Pisanelli and S Vitellozzi (2025), "Beyond Time: Unveiling the Invisible Burden of Mental Load" CEPR Discussion Paper No. 20269. CEPR Press, Paris & London. ISSN 0265-8003 

This paper introduces a novel, scalable methodology to measure individual perceptions of gaps in mental load—the cognitive and emotional burden associated with organizing household and childcare tasks—within heterosexual couples. Using original data from the TIMES Observatory in Italy, the study combines time-use diaries with new survey indicators to quantify cognitive labor, emotional fatigue, and the spillover of mental load into the workplace. Results reveal systematic gender asymmetries: women are significantly more likely than men to bear organizational responsibility for domestic tasks, report lower satisfaction with this division, and experience higher emotional fatigue. These burdens are underestimated by their partners. The effects are particularly pronounced among college-educated and employed women, who also report greater spillovers of family responsibilities than men during paid work hours. The perceived responsibility for managing family activities is more strongly associated with within-couple gaps in time use than with the absolute time spent on their execution, underscoring the relational and conflictual nature of mental load.

4. Pisanelli, E. "Measuring domestic violence: Individual attitudes and time use within the household" Working paper available at arXiv:2511.01473. DOI:10.48550/arXiv.2511.01473

Time spent with a partner is widely interpreted in time-use research as revealed-preference evidence of cooperation and relationship quality. This paper argues that co-presence can instead reflect constrained choice when norms legitimize monitoring and control. Using original matched-couple data from Italy, I link vignette-based measures of intimate partner violence (IPV)-tolerant normative orientations to high-frequency time-use diaries completed by both partners. The diaries record not only activities but also who is present, allowing me to measure women’s autonomy-relevant margins—time spent alone versus with the partner, time outside the home, and solo childcare—alongside specialization in unpaid work and childcare. Across couple-level models and actor–partner interdependence specifications, IPV-tolerant orientations exhibit positive assortative matching within couples and are associated with more gendered divisions of unpaid work and childcare. Consistent with a coercive-control interpretation, the associations with reduced autonomy are stronger when women have low economic independence. The findings highlight an interpretation problem in time-use data: co-presence is behaviorally ambiguous and can coexist with hierarchical household organization shaped by harmful norms.

5. Pisanelli, E.; Schram, A. "Even Biased AI Recommendations May Improve Job Candidate Selection"

Organizations increasingly use artificial intelligence (AI) to assess job candidates. This has raised widespread concerns about potential reinforcement of the discrimination of certain groups. While recognizing this possible downside, this paper explores a different consequence that the use of AI in candidate evaluation may have for hiring bias and experimentally tests the theoretical predictions. Our model is based on the literature on information cascades. Two types of candidates are considered, one is randomly chosen to have a generic (but unknown to the employers) advantage in productivity. Each candidate also has a randomly assigned private productivity component. Employers sequentially choose between two candidates (one of each type), receiving independent signals of candidates’ productivity components and in some environments, observing previous employers’ choices. We introduce AI’s information on private productivity, which can be (i) none, (ii) an unbiased signal coming from AI, or (iii) a knowingly biased signal coming from AI. Theoretical and experimental results show that (ii) and (iii) can break information cascades with positive probability. This introduces a channel through which AI may improve recruiters’ decisionmaking even when providing biased information.

Projects