Giovanni Mellace

Limited Monotonicity and the Combined Compliers LATE Abstract. We consider endogenous binary treatment with multiple binary instruments. We propose a novel limited monotonicity (LiM) assumption that is generally weaker than alternative monotonicity assumptions in the literature. We define and identify (under LiM) the combined compliers local average treatment effect (CC-LATE), which is arguably a more policy-relevant parameter than the weighted average of LATEs identified by two-stage least squares (TSLS), and is identified under more general conditions. Estimating the CC-LATE is trivial, equivalent to running TSLS with one constructed instrument on a subsample. We use our CC-LATE to empirically assess how knowledge of HIV status influences protective behaviors.

A new way to estimate causal effects when using several IVs. Just Accepted new paper by Nadja van’t Hoff, Arthur Lewbel, and Giovanni Mellace zurl.co/XZvKK

Causal Inference for Qualitative Outcomes Causal inference methods such as instrumental variables, regression discontinuity, and difference-in-differences are widely used to identify and estimate treatment effects. However, when outcomes are ...

🎉 Thrilled to share that my paper with Riccardo Di Francesco has been accepted at Economics Letters!
🔗 arxiv.org/abs/2502.11691

Gender Differences in Healthcare Utilisation - Evidence from Unexpected Adverse Health Shocks This paper is the first to provide causal evidence of gender differences in healthcare utilisation to better understand the male-female health-survival paradox, where women live longer but experience ...

Excited to share our new paper on arXiv! 📢
With my amazing colleagues Seetha Menon and Nadja van 't Hoff, we show that—contrary to common belief—men use more healthcare than women when faced with an unexpected health shock.

🔗 arxiv.org/abs/2509.01310

Excited to present at #iHEA2025 in Bali 🌴:
📄 Nudging Nutrition: Lessons From the Danish "Fat Tax"
📄 Sex Differences in Healthcare Utilisation: Evidence from Unexpected Adverse Health Shocks
#IHEA205

Looking forward to presenting in Torino at #IAAE2025 our Shapley Instruments paper in a super interesting session. Join us on Wednesday if you attend the conference

It's always a great time visiting my coauthor @fedeabides.bsky.social in Siena. This time I also presented our brand new paper Shapley Instruments joint with Nadja van't Hoff and Zsolt Sandor.

I had such a great time at #espe2025 in Napoli. Very interesting presentations and outstanding social events in an amazing location.

🎉 Excited to share that our paper "Limited Monotonicity and the Combined Compliers LATE" (with Nadja van’t Hoff and Arthur Lewbel) has been accepted for publication in the Review of Economics and Statistics!
📄 fmwww.bc.edu/EC-P/wp1059....

#ICEEE2025

Looking forward to presenting my paper with my great coauthor @rokspruk.bsky.social in Palermo on Friday. We evaluate the huge investment in public health and education that followed the revolution finding long lasting effects.

PhD summer school PhD Course: Historical Perspectives on Current Economics Issues: Big Data and Applications HEDG and Department of Economics, University of Southern Denmark 12–22 August 2024

🚨 PhD course @sdu.dk this August!
"Historical Perspectives on Current Econ Issues: Big Data & Applications"
📅 Aug 11–22 | Odense
👨‍🏫 Christopher Meissner (UC Davis)
Apply by July 1 www.sdu.dk/en/forskning...

Wow!

🚨🚨🚨PhD summer school at the University of Southern Denmark 🚨🚨🚨What can we learn about tariffs and trade from history? Chris Meisner, our keynote speaker, is one of the most renowned experts on historical trade and policy. + Learn ML + NLP tools in a hands-on-approach. Join us 11-22 August 2025!

Looking for a salaried PhD position? Join my colleague in a cutting-edge research project at SDU, in collaboration with MIT, to explore the transformative impact of disability reforms with register data!
Apply here: fa-eosd-saasfaprod1.fa.ocs.oraclecloud.com/hcmUI/Candid...

Causal Inference for Qualitative Outcomes Causal inference methods such as instrumental variables, regression discontinuity, and difference-in-differences are widely used to estimate treatment effects. However, their application to qualitativ...

Is your outcome a qualitative variable, but you still want to conduct causal inference?
Check out our new paper Causal Inference for Qualitative Outcomes!
We provide an open-source R package,causalQual , for easy implementation:
🔗 github.com/riccardo-df/...

📄 arxiv.org/abs/2502.11691

Econometrics and Data Science

We’re looking for a motivated researcher to apply for a Marie Skłodowska-Curie postdoc with our Econometrics & Data Science group at SDU!

Focus: Causal Inference, Machine Learning, Big Data
Full support for promising projects

More info & apply:
www.sdu.dk/en/om-sdu/in...

🚨NEW WP OUT!🚨 Our working paper with Claudio Detotto & Carlo Caporali on the causal impact of terrorism on voter turnout during Italy's Years of Lead is out! Comments and suggestions welcome! Link here: crenos.unica.it/crenos/publi...

Our latest research with Toman Barsbai and Christoph Winter is now forthcoming in JOLE! 📄

We explore how labor market conditions at arrival shape migrants’ integration, showing that arriving during recessions pushes migrants into lower-paying jobs—a persistent effect lasting over 10 years.

We are looking for a full professor in economics with Focus on Big Data and Large Language Models in Social Data Science. Deadline: February 17, 2025
fa-eosd-saasfaprod1.fa.ocs.oraclecloud.com/hcmUI/Candid...

New WP 🚨

1. Recipe to write estimators as weighted outcomes
2. Double ML and causal forests as weighting estimators
3. Plug&play classic covariate balancing checks
4. Explains why Causal ML fails to find an effect of 1 with noiseless outcome Y = 1 + D
5. More fun facts
arxiv.org/abs/2411.11559

Looking to hire a talented and dedicated econometrician? I highly recommend Nadja van't Hoff. We collaborated with
Arthur Lewbel on "LiM and the CC-LATE", R&R at
the REStat. In her JMP, she develops novel identification results for settings with non-binary, ordered treatments. nadjavanthoff.com

The inclusive Synthetic Control Method We introduce the inclusive synthetic control method (iSCM), a modification of synthetic control methods that includes units in the donor pool potentially affected, directly or indirectly, by an interv...

Roberta Di Stefano and I have updated our paper, "The Inclusive Synthetic Control Method." The new version, now incorporating the penalized SCM in our application, is available here: arxiv.org/abs/2403.17624.

The US elections in a nutshell 🤦‍♂️https://www.dailymail.co.uk/video/news/video-3306573/Video-time-Puerto-Rican-Trump-voter-reveals-shocking-reason.html

Sciences Po has junior positions in econometrics and in health/environmental this year, please apply, and please RT!
econjobmarket.org/positions/10...

We are highering a tenure-track Assistant Professor in Macroeconomics, deadline December 1, 2024. eur03.safelinks.protection.outlook.com?url=https%3A...

Please like and share. Help us to spread the word in these post-Twitter times! 🙏

The gray zone: How not imposing a strict lockdown at the beginning of a pandemic can cost many lives The public debate on the effectiveness of lockdown measures is far from being settled. We estimate the impact of not having implemented a strict lockd…

I'm happy to share that our paper with @giovannimellace.bsky.social , Roberta Di Stefano, and Silvia Tiezzi has been accepted in Labour Economics. www.sciencedirect.com/science/arti...

🚨New language model for historical occupations🚨
OccCANINE is a new tool, which turns occupational descriptions (in all their pesky variety) into standard HISCO codes. Accurate, multilingual and fast. Developed by
Christian M. Dahl, Torben Johansen, and myself (1/12)

The inclusive Synthetic Control Method We introduce the inclusive synthetic control method (iSCM), a modification of synthetic control type methods that allows the inclusion of units potentially affected directly or indirectly by an...

Roberta Di Stefano and I have substantially updated our paper the "inclusive Synthetic Control Method" (available here arxiv.org/abs/2403.17624). Our iSCM allows to include units affected directly or in indirectly by the treatment in the donor pool and it works with most SC-type estimators.

Very happy to contribute to this great initiative to #SupportUkraine
When? April 11th
What? 2h of fun understanding how (easy) DoubleML and Causal Forests of @VC31415 and @Susan_Athey work by replicating R output in few lines of code. #RStats
How? Sign up here t.ly/3weFr