🚨New Working Paper🚨
Was Javert right to be suspicious? Unpacking treatment effect heterogeneity of alternative sentences on time-to-recidivism in Brazil
by Acerenza (), Possebom and Sant’Anna (
@pedrohcgs
)
Paper:
💡Econometrics Thread💡Today, I will talk very briefly about a few recent methodological papers that I think are super useful to applied researchers.
Basically, below, you will find some new tools that may help you to answer relevant empirical questions.
🎉Career/Life Update🎉I am super happy to tell everyone that I will start as an assistant professor at FGV EESP in May 2022. I learned a lot as a student there and I am looking forward to going back.
I am also excited to move back to São Paulo, a city that I have always loved.
Today, I’m moving back to São Paulo, where I’m gonna start as an assistant professor at
@EconFGVSP
.
I’m very excited about this new stage in my academic career and I’m looking forward to work in-person with my colleagues.
💡Econometrics Thread💡Marginal Treatment Effects (MTE)
Today, I will not talk about one particular paper. I will explain the basics of MTE and will mention the papers that I believe to be important. As my wife says, I like to preach and spread the Gospel of MTE.
#EconTwitter
, I need your help. I love econometric books with coding examples, such as:
- The Mixtape by
@causalinf
- The Effect by
@nickchk
- Brave and True
@MatheusFacure
.
But they are all focused on X-section and panel.
Do you recommend anything focusing on time series?
I am an international student and I want to ask all the professors to email their university's administration to help us with this immigration issue. We may lose our right to stay in the US. We came here with dreams and hope to build a life and a career here.
This is bad. ICE just told students here on student visas that if their school is going online-only this fall, the students must depart the United States and cannot remain through the fall semester.
🚨Publication Alert🚨"Crime and Mismeasured Punishment: Marginal Treatment Effect with Misclassification" by Vitor Possebom (2023, )
I am so happy that my JMP was accepted at the REStat!
I am thankful to so many people who helped me through this journey.
🚨Acceptance Announcement🚨 I am truly thrilled to announce that my paper "Identifying Marginal Treatment Effects in the Presence of Sample Selection" with my amazing co-authors
@OBartalotti
and Désiré Kédagni has been accepted for publication at the Journal of Econometrics!
💡Metric Thread💡"OLS is complicated, mate! Proceed with caution!"
I love papers that look at OLS, a centuries-old tool, and still find something we did not know about it. They are surprising and very relevant for applied work. Here, I will list a few recent ones.
#EconTwitter
💡Short Econometrics Thread💡Today, I will talk very briefly about a few recent papers expanding the synthetic control method (SCM). I will talk about three working papers: Abadie and L'Hour (2019), Grossi, Lattarulo, Mariani, Mattei and Oner (2020) and Cao and Dowd (2019).
💡Short Econometrics Thread💡Multiple Hypothesis Testing in Experimental Economics by List, Shaikh and Xu (2019, )
Multiple Hypothesis Testing (MHT) is the kind of comment that econometricians love to make and applied researchers hate to hear.
REStat has been one of my favorite journals for awhile now. It publishes papers that I like to describe as midfield econometrics, connecting theoretical and applied work. I’m deeply honored of publishing a paper there.
LIV is biased for the MTE when the treatment is misclassified. But it is still possible to set ID the MTE. Just Accepted new paper by Vitor Possebom (
@PossebomVitor
)
💡Metrics Thread💡Conditional Average Treatment Effects (CATE)
Most people know the definition of the CATE: it is the ATE for a group of people that share an observable characteristic x, i.e., CATE = E[Y1 - Y0| X = x].
But why is it so cool?
💡Econometrics Thread💡Marginal Treatment Effects from a Propensity Score Perspective by Xiang Zhou and Yu Xie (2019, JPE, ).
Zhou and Xie (ZX) wrote an easy-to-read research note on the Marginal Treatment Effect (MTE).
💡Econometrics Tread💡Identification, Inference and Sensitivity Analysis for Causal Mediation Effects by Imai, Keele and Yamamoto (IKY, 2010, )
Mediation analysis decomposes a treatment effect in different causal mechanisms.
💡Econometrics Thread💡”Cherry Picking with Synthetic Controls” by Ferman, Pinto and Possebom (2020, ).
It was a long journey, but we have finally published our paper at
@JPAM_DC
. I’m really happy to see it on the journal’s website.
I am very honored to receive such a significant prize.
I thank my advisors - Ed Vytlacil, Xiaohong Chen,
@narita_yusuke
, and
@john_eric
.
I also thank my co-authors:
@OBartalotti
and Désiré Kédagni.
Moreover, I thank everyone who helped me during my Ph.D. journey.
@PossebomVitor
, a former masters student and current faculty member, won the Best Ph.D. Thesis Prize at
@anpec_br
’s meeting!
His thesis has two chapters:
1) MTE with Misclassified Treatment:
2) MTE with Sample Selection:
💡Econometric Thread💡"Identification Strategies and Estimators are not the same thing!"
The thread is motivated by the recent and excellent paper "When is TSLS actually LATE?" by
Blandhol, Bonney, Mogstad and Torgovitsky (BBMT, ).
🚨Working Paper Alert🚨"Crime and Mismeasured Punishment: Marginal Treatment Effect with Misclassification" by Vitor Possebom (2021, )
I am super excited about this project and I want to share its theoretical part. The empirical part is coming soon!
💡Econometrics Thread: 💡Going beyond average effects with Diff-in-Diff. Recently, DiD methods became very popular and many excellent papers showed up! But, today, I want to talk about two papers that are not that recent, but are super interesting!
Brazil will celebrate its Independence Day on September 7th. I was planning to write a celebratory thread on 09/07, but I decided to anticipate it!
So, it is time for a thread about papers using Brazilian data! I will mention one paper per econometric method.
I am a fan of
@marcelamellos
’s work. Besides her amazing JMP, she has two WPs with
@joaomgarcia
and Latham-Proença.
Since I like early-childhood development, I will highlight their paper about the impact of free child care on mothers’ careers: .
I just prepared a syllabus for a new course that I will teach this year: Economics of Crime for undergraduate students!
I am super excited about it and we will cover many recent papers discussing relevant policy questions! These policies are super hot topics in the press!
💡Paper of the Week💡
Have you tried to estimate the average treatment effect using a parametric propensity score (ps) and wondered: "What if my model is misspecified?"
@pedrohcgs
and Song can help you. Check their recent JoE paper: .
O Pichetti foi meu professor de time series na graduação e hoje é meu colega. Ele entende muito de econometria, macroeconomia e finanças.
Ele vai contribuir muito para o país no BC. Estou muito feliz por ele e pelo Brasil. Tenho certeza que ele fará um trabalho sensacional.
O professor da graduação em Economia da FGV EESP, Paulo Picchetti, foi indicado para fazer parte da diretoria do Banco Central. O professor é especialista em econometria, análise de ciclos econômicos e índices de preços. Para saber + acesse:
💡Short Econometric Thread💡"When Is Parallel Trends Sensitive to Functional Form?" by
@jondr44
and
@pedrohcgs
(2020, )
This morning, I watched Jonathan's presentation at the Chamberlein Seminar () about his paper with Pedro.
As an end of the year thread, I would like to highlight some of the 2021 papers that made me think and reflect. They are not in any special order and they reflect my empirical and theoretical interests.
So here is my list!
Post your favorite papers here too!
TWITTER SUPPORTS MASSACRES: In the last weeks, Brazil had two mass killings in schools.
The Brazilian government asked
@Twitter
to remove accounts that support the murders.
@Twitter
refused to help the Brazilian government.
On Tuesday, the 42nd Meeting of Brazilian Econometric Society is gonna start ().
I’ll take this opportunity to highlight some brilliant work from young Brazilian researchers. I’ll focus on people I know personally and whose research I find inspiring.
This week, I got my copy of the amazing “Causal Inference: The Mixtape” by
@causalinf
. His chapter on Synthetic Controls is extremely clear, explaining the most important aspects of the method.
I rarely say anything about heated debates on
#EconTwitter
, but I want to share my opinion on this topic because it changed a lot over time.
Brazilian law mandates every Econ undergrad to take a course about History of Economic Thought and about Brazilian Economic Formation.
Almost everything worth reading in economics is covered at length in grad level textbooks.
Nobody learns calculus by reading De analysi per aequationes numero terminorum infinitas.
If you want to read original classic texts there's a discipline for you: comparative literature.
💡Econometric Thread💡Equivalence Testing for RDD and DiD
In applied microeconometrics, it is common to indirectly test the identifying assumptions behind your identification strategy.
For example, we run placebo tests in RDD and we test for pre-trends in DiD.
🚨Updated Working Paper🚨"Crime and Mismeasured Punishment: Marginal Treatment Effect with Misclassification" by Vitor Possebom (2021, )
As promised, I finished my empirical application and I will, now, write a thread about the whole paper!
And
@EconFGVSP
is looking for new assistant professors! You can find our ad on .
My first semester here has been amazing! We are a very friendly department with a nurturing environment. One thing that I love here is that we all have lunches together!
Yesterday, I had the great pleasure of visiting
@EPGE_FGV
. Thank you for inviting me,
@barsanetti
and Marcelo.
It was awesome to meet their young APs,
@smathecon
,
@MarcosSalgado46
and Andrea.
I guess that EPGE is likely to be the Econ department with the best view in the world.
🔥Hot Take🔥 I like the long papers that are so common in economics. They usually have detailed discussion that allow me to understand the literature, the data, the institutional background and the methodology.
Yesterday, I worked in polling station for the Brazilian elections. The public school where I was voting was very colorful and well-kept. As soon as I arrived there, I thought about this excellent paper by
@Nicolas_Ajz
and Durante: .
Are you concerned with measurement error? Do you wanna know how partial identification can help you address this problem? This thread is for you! (
@dlmillimet
, I was inspired by your blog post!) 1/14
We are hiring!
@EconFGVSP
is an amazing place!
We have a nurturing and rich research environment! And that’s awesome for APs like me,
@AraujoCRRafael
,
@finamor_lucas
, Pierluca and Henrique!
But the thing that I love the most is that we are an enormously friendly place!
We are hiring!
@EconFGVSP
invites applications for full-time professor positions. We seek applicants working in any field of economics. Applicants working in macroeconomics and microeconomic theory are encouraged to apply.
Apply here:
I basically only post about two topics: Metrics and Brazilian food. Today, I’m a happy Brazilian. ☺️ And I’m glad I discovered
@FaviFoods
! And I loved their “Thank you, Lindezas!” card.
🚨New Working Paper🚨"Probability of Causation with Sample Selection: A Reanalysis of the Impacts of Jóvenes en Acción on Formality" by Possebom and Riva (2022, )
After so many years,
@friva_
and I finally worked together on a paper. He is amazing!
Partial Identification strategies are key for any work in Econometrics! I would love to see them becoming more common in applied work! We are all used to standard errors as measures of sample uncertainty and we would never believe in a result before seeing the CI!
How to effectively teach kids how to read?
This is a big question whose answer is context specific, but I loved this AEJ Policy paper: .
The authors find that the phonic teaching method works better than the whole language approach.
It was evaluated in a very impressive piece of economic research by LSE
They found huge positive effects. At very low costs. The same magnitude as the famous gold-standard "reducing class sizes" STAR intervention... but much much cheaper.
💡Short Metrics Thread💡I always like to read papers that are not directly related to my research and this week was particularly fruitful in this area.
Below, I list this week's papers. There is RDD, DiD and Randomization Inference. And a non-econ surprise at the end.
This afternoon I decided to learn a little bit about gganimate and illustrated the convergence in distribution of the MLE estimator for an MA(1) process.
I was surprised about how easy gganimate is! It is ncredibly intuitive.
💡Random Thoughts💡I was thinking about what I like to see in an applied paper using the synthetic control method (SCM). What makes me like a synthetic control paper? My opinion is based on subjective criteria. You may disagree and I would love to hear your opinion on this issue.
Eu gostaria de agradecer o convite do
@ppge_ufpb
e do Vinicius. Foi um prazer conhecer os professores e os alunos da UFPB. Tem muita pesquisa legal sendo feita aqui!
Today is teachers’ day in Brazil. To celebrate this day, I would like to tweet about a econometrics paper whose empirical application is about teacher value added: .
Kwon (2021) proposes a shrinkage estimator that does not rely on normality assumptions.
Keeping with the tradition of highlighting Brazilian economists before our Independence Day, we have another great scholar: Luciene Pereira ().
Her research applies micro-level evidence to understand macroeconomic outcomes over time and across economies.
One of my favorite conferences is the
@SBEconometria
's meeting and it will happen next week!
Here, I will list the amazing sessions I plan to attend. But I do not promise I will be able to do so.
💡Econometrics Thread💡 Case-control control
Recently, I read a couple of papers about case-control studies, a sampling method that is common in epidemiological studies.
#epitwitter
And I have been thinking about some of the techniques I learned in those papers.
Super cool thread on the advantages of using the wild bootstrap when you have only a few clusters. If you want to have a deeper understanding about the asymptotic theory behind the wild bootstrap, check this recent paper by Canay, Santos and Shaikh:
How biased are clustered SEs with 'few' clusters? A simulation illustrating this. DGP is y~x, 50 clusters, x is normal, true beta is 0.5. Plot of 1000 sims, beta estimate +95% CIs for each. Red = we did not cover true beta. Std SEs no good, clustered SEs yield ~95% coverage (1/6)
@carolcaetanoUGA
I really like Introduction to Econometrics with R: . It is a free online book based on Stock and Watson (same chapters and content, but with R code through the text).
💡Econometrics Thread💡 “A correction for regression discontinuity with group specific mismeasurement of the running variable” by
@OBartalotti
, Brummet and Dieterle (2020, BBD from now on). Forthcoming at JBES: . Ungated version:
💡Econometric Thread💡"Marginal Treatment Effects with Misclassified Treatment" by Santiago Acerenza, Kyunghoon Ban and Désiré Kédagni (ABK, 2021, )
This thread will be longer than usual and (hopefully) more detailed.
💡Applied Thread💡 “Cash Transfers and Formal Labor Markets: Evidence from Brazil” by Gerard, Naritomi and Silva (2021, )
Yesterday, I read this amazing paper about the positive impact of a CCT program on formal labor markets.
Luis (
@lafalvarez
) is a very bright and productive young econometrician. You should follow him to know more about his amazing work.
Today, I want to highlight some of his current working papers. 🧵
We are happy that Luis Alvarez (
@lafalvarez
), our master alumnus and current Post-Doc, will start his career as an Assistant Professor at FEA-USP (
@FEAUSPcomunica
).
He is an econometrician. To learn more about his work, check out his website: .
In my opinion, an equilibrium where scholars from different countries communicate in a single language boosts cooperation and increases research quality. Today, English established this type of equilibrium.
💡Econometric Thread💡"Asymptotically Uniform Tests After Consistent Model Selection in the Linear Regression Model" by Adam McCloskey (2020, )
McCloskey wants to test the effect of X on Y when we are unsure about which covariates Z should be included.
Hey,
#EconTwitter
and
#epitwitter
! I will start teaching an introductory causal inference course next week and I would like to recommend a few accessible podcasts to my students.
I know that
@ProbCausation
and
@casualinfer
are great options. Any other recommendations?
💡Short Econometrics Thread💡"An Introduction to Proximal Causal Learning" by Eric Tchetgen Tchetgen, Andrew Ying, Yifan Cui, Xu Shi and Wang Miao (2021, )
Tchetgen Tchetgen presented his super cool paper today at Yale's Political Science Department.
Today, Isaiah Andrews won the Clark Medal. I had the pleasure of meeting him when he visited Yale a few years ago. He is such a kind soul, asking thoughtful questions in all seminars. His research is also awesome, tackling econometric problems that matter for applied research.
If you want to drink from a peaceful DiD water fountain instead of from a DiD fire hydrant, you can start with
@Andrew___Baker
’s thread about his new paper with Larcker and Wang. It is an applied guide about recent advances in the DiD literature.
Today, my hometown, São Paulo, is celebrating its 468th anniversary!
To celebrate this date, I’ll share two papers that uses data from São Paulo city!
What are your favorite papers using data from the coolest city in Brazil? 😜
💡Econometrics Thread💡"Accounting for Heterogeneous Returns in Sequential Schooling Decisions" by
@gema_zamarro
(2010, ).
Today, I will write about a paper that I discovered here. So the thread is also an example about why
#EconTwitter
is cool.
This macro course was taught using this textbook: .
It covers many ways to think about the macroeconomy, historically contextualizes each school and contrasts all schools, helping us to see how our understanding of Economics changed over time.
As I already said,
@LaceaLames2021
has many amazing sessions. On Friday, there will be a cool session about Marginal Treatment Effects at 9:45 ET You should come! There will be four super interesting presentations! (Obviously, my opinion is biased. Hehe)
And you may say I am a broken record, but
@analisereal
's and
@chadhazlett
's paper is amazing: . I really think it should become a standard robustness exercise in any paper using CIA via OLS.
Check Carlo's thread about it:
(1/10) Happy to announce that my paper with
@chadhazlett
is officially out on JRSS-B! ()
In this paper, we develop new sensitivity analyses tools to precisely quantify how strong confounders need to be to overturn your research conclusions.
#EconTwitter
, I have a question. It is for a friend.😅This friend is writing a letter replying to each referee's comments.
Is it common that the response letter is as long as the manuscript?
In his defense, he copied each referee's comment in the response letter.
@PhDemetri
Since the usual regression model is only a model about the mean of Y and does not require any normality assumption, what are the advantages of teaching it with more assumptions than needed?
I am so happy that the Sao Paulo School of Economics (
@EconFGVSP
) joined Twitter today!
We will post about our research seminars, papers published by our faculty members and students, and much more.
Today, I want to highlight the accomplishments of
@EconFGVSP
's students and highlight their papers at
@SBEconometria
's meeting.
I will try to tag all students, but, if I do not find someone, feel free to tag yourself.
More importantly, if I miss someone, let me know in the DMs!
So, currently, I still hate my History of Economic Thought. It was about how some really old people thought about economics in a outdated way. But I would love a course that uses historical thinkers as a way to explain how Economics became what Economics is today.
The students at
@EconFGVSP
are amazing and I love working with them.
In my first year here, I was very impressed with their theses and learned a lot about various topics, ranging from agricultural economics to banking competition.
If you are interested in doing an Econ Ph.D. in Brazil and is currently residing in another country, check our Application Guidelines: .
Our department is very friendly and interactive. You can learn more about it here: .
Sant’anna e Callaway é o meu paper favorito de DiD dos últimos anos. Além disso, o
@pedrohcgs
apresenta muito bem e é super didático. No seminário dessa semana, você vai aprender muito.
Esta quarta teremos o ultimo seminário da temporada 2020 com
@pedrohcgs
apresentando seu paper que é apontado como um dos avanços recentes em micro aplicada:
@claudferraz
@fgv_clear
I am preparing my time series lecture slides and I was looking for examples of time series with trends. Honestly, I find it impossible to not feel optimistic when I see the time series for extreme poverty and child mortality.
Humanity achieved so much!
@OurWorldInData
💡Updated Working Paper💡 Crime and Mismeasured Punishment: Marginal Treatment Effect with Misclassification
After excellent reports by the editor and referees, I have updated my paper about the MTE with a misclassified treatment variable.
That’s what
@Undercoverhist
does in her threads. That’s why I loved reading about the debates between Imbens, Heckman and Deaton or between Frisch, Fisher and Keynes or between the NBER and the Cowles Commission. That’s also the research done by
@arthurbnetto
.
Hi,
#EconTwitter
. Some of my students want to publish their senior thesis. I found two lists of journals that publish undergraduate research: and . Some of them are multidisciplinary and some of them are specific to economics.
Continuous DiD has always fascinated me and it seemed more complicated than applied papers were actually doing. Callaway,
@agoodmanbacon
and
@pedrohcgs
address all the complicated details in this super clear and insightful paper.