$$ \newcommand{\N}{\mathbb{N}} \newcommand{\R}{\mathbb{R}} \newcommand{\E}{\mathbb{E}} \renewcommand{\Pr}{\mathbb{P}} \newcommand{\Var}{\mathrm{Var}} \newcommand{\Cov}{\mathrm{Cov}} \newcommand{\Corr}{\mathrm{Corr}} \newcommand{\plim}{\operatorname*{plim}} \newcommand{\symbf}[1]{\boldsymbol{#1}} \newcommand{\iid}{\overset{\text{iid}}{\sim}} $$
Lecture 0: Class Setup
Raul Riva
FGV, Brazilian School of Economics and Finance (FGV EPGE)
October 2026
Econometrics I (Part I)

Course Information

We will use GitHub for course materials, assignments, and so on.

Course Website

QR code linking to the course website

https://epge-classes.github.io/econometrics-fgv-2026/

Why Should You Take This Class Seriously?

  • Most students will engage in empirical research across diverse fields of Economics:
    • Labor, Development, Urban, Finance, Health, IO, even Macro…!
    • Obviously important if you want to do Econometrics
  • So far, most of what you learned is about cross-sectional data and OLS
  • This class: Time Series, Causal Inference, and Panel Data
  • Time Series data is your bread and butter in Macro, Finance, and many other fields
  • Panel Data methods (fixed effects, diff-in-diff, …) are everywhere in empirical research
  • Needless to say why Causal Inference is important, right? 👀
  • We will also cover a bit of non-parametric methods and bootstrap techniques

Learning Goals (by Week 5)

After the first half of the course, you should be able to:

  1. Understand and implement non-parametric kernel-based methods
  2. Apply bootstrap methods for i.i.d. data
  3. Analyze time series data:
    • Concepts: stationarity, ergodicity, persistence, autocorrelation, …
    • Estimation and forecasting with ARMA models
    • Asymptotic distribution of the OLS estimator under autocorrelation
  4. Implement GMM estimation under heteroskedasticity and autocorrelation:
    • Super relevant for Panel Data methods
    • Also very useful to estimate structural models

Tentative Weekly Topics

  • Week 1: Non-parametric estimation + Bootstrap for i.i.d. data
  • Week 2:
    • The structure of time series data, stationarity, ergodicity, persistence, autocorrelation, …
    • The Lag Operator, AR and MA processes
  • Week 3:
    • ARMA Models and Wold’s Decomposition Theorem
    • Estimation of ARMA Models
  • Week 4:
    • LLN and CLTs for dependent data – focus on using the results rather than proving every single little thing
    • Intro to GMM
  • Week 5:
    • GMM asymptotic theory + HAC estimation
    • GMM and HAC estimation + numerical pitfalls

Note: subject to change.

Books? Slides?

  • I will provide slides for the lectures
  • These slides will be mostly based on two books + a famous reference here and there:
    • Econometrics, by Bruce Hansen (hello, old friend!)
    • Time Series Analysis, by James Hamilton (very comprehensive, also very long)
    • Stochastic Limit Theory, by James Davidson (seriously dark magic)
  • I will highlight the recommended readings, and reading the books is mandatory
  • Slides are not a substitute for reading the books
  • Slides available on GitHub just before class

Evaluation and Grades

  • Grades for the first and second half of the class are independent
  • Final grade = 0.5 \(\times\) (Grade on Part I) + 0.5 \(\times\) (Grade on Part II)
  • Problem sets are worth 10% of my half of the course; exams are worth the remaining 90%

Problem Sets

  • Groups of up to 3 students
  • Late solutions are not accepted
  • Both theoretical and empirical questions
  • Solutions must be submitted in PDF format through GitHub
  • We will use GitHub Classroom to manage problem set submissions – see problem_sets
  • There will be one private repository for each group and each problem set
  • Deliver both a report answering the questions and the code used to generate the results
  • Feel free to use any civilized programming language (Python, R, Julia, Matlab, …)
  • No, you cannot use Stata or pre-packaged solutions – be honest and sensible

AI Policy

You can, and you should, use AI tools in this class. But be smart about it.

  • ● Use AI to understand the material and to help you write code
  • ● Do not use it to blindly write your reports
  • The written exam may ask questions about the empirical exercises
  • No AI help during the exam – and exams are 90% of your grade
  • Some problems might explicitly ask you to use AI tools

Attendance

  • It is not mandatory at all – you are adults
  • However, if you come, you have to be on time
  • 15-minute grace period – after that, you will not be allowed to enter the classroom

 

Questions?

One more thing…

This is really important!

  • I need ongoing feedback
  • Is there something great/terrible about the class? Tell me!
  • Don’t be shy – negative feedback will never be punished in any way
  • If anything, on the contrary!