Version 220822 Syllabus Econometrics I, 310153-M-3, Fall 2022 Lecturer Ben Vollaard, vollaard@uvt.nl Teaching assistant Antonio Laplana, laplana@uvt.nl Office hours Upon appointment. Canvas course page See: https://tilburguniversity.instructure.com/courses/10935 Prerequisites We assume that you have an understanding of basic econometrics; successful completion of an introductory course in econometrics is required. Please contact the lecturer before the start of the course if you want to follow this course, but do not feel that you have a sufficient grasp of statistics and regression analysis. Textbook We use a textbook that just came out: Nick Huntington-Klein, 2022, The Effect: An Introduction to Research Design and Causality, Chapman & Hall (routledge.com covers Europe, use code JML20 to get a 20 percent discount), also available as E-book. A free Bookdown version is available at: theeffectbook.net (with better font and layout compared to the published version). To run the R (or Stata/Python) code examples in the book, you need the causaldata package, which contains the example data for most of the code chunks and can be installed using install.packages('causaldata') in R. We occasionally fall back on basic econometrics as worked out in a textbook like Jeffrey Wooldridge, Introductory Econometrics (used in the BSc. Econometrics course). Software We work with in this course, free open source statistical software. Our user interface for R is RStudio. Other MSc courses, including Labor Economics, also use R. I strongly recommend using RStudio Cloud . https://rstudio.cloud/ With R Studio Cloud, on any computer with internet access, including university computers, you can just go to the Version 220822 cloud and work on your assignment. You can also work on an assignment as a team (for more on student teams, see below). In the past, several students ran into problems with operating systems that do not always co-operate with R, especially Mac OS. In addition, several students got stuck on computer assignments because they unknowingly worked with old versions of R or of specific packages. In short, working in the cloud is convenient and a big time saver. We do not have a university license for RStudio Cloud, but the fee plans are very reasonable. Maybe the ‘Free plan’ is already sufficient, given that it provides 25 hours per month. If you do need additional time, then you may want to consider the ‘Plus plan’ (75 hours per month for $5 per month). (Only) if you want to run R and RStudio locally, then download and install both programs before the start of the course. For R, see https://cran.r-project.org/bin/windows/, for RStudio Desktop see https://www.rstudio.com/products/rstudio/download/ Please note: if you have little or no experience with R, follow the intro-to-R course in the week prior to the first lecture. Otherwise, you will be in serious trouble; completing the weekly computer assignments will then take an inordinate amount of time. For more information: see the announcement on the Canvas course page. Description of the course This is a course in research design. A research design is the qualitative story that is going to give you (quasi)-experimental variation in the variable that you want to manipulate. This is independent of whether you are doing micro or macroeconomics. Experimental variation allows for a causal interpretation of the results of your quantitative analysis, which is our aim. We focus on questions related to policy evaluation. Put differently: a research design is like a mold in which you pour the data (see figure). Obviously, before you do anything, you start with making a mold. So, rather than simply regressing variables on each other, you learn to think about the method of your study first. After introducing the fundamental problem of causal inference, we discuss five commonly used research designs: selection on observables, randomized trial, regression discontinuity, differencein-difference, instrumental variable. Version 220822 The course should help you to conduct meaningful and creative empirical work by yourself, by learning to recognize research designs in the day to day world. In addition, this course should help you to assess the quality of empirical work from others. Course planning The course starts in the week of August 28, 2022. The course runs for five weeks in a row. Please note that in the preceding week, i.e. the week starting Monday August 22, you should follow the above-mentioned intro-to-R course. Teamwork The two weekly computer assignments (more about this later) should be made in groups of three (not two, not four). You are invited to form groups on Canvas; we will assign all other students to groups of three in the first week of the course. Weekly schedule The weekly schedule consists of three sessions, as shown below: 1. Reading, the occasional clip (Monday), lecture (Tuesdays, generally at 08.45 am) The theory related to that week’s topic is presented in the lecture on campus. You are expected to do the reading before attending the lecture, as it builds upon the reading. Occasionally, accompanying knowledge clips are posted on the lecturer’s YouTube channel. A knowledge clip is a short video of a couple of minutes that covers a single, specific subject. Please note any room and time changes on your schedule. 2. Tutorial session: Q&A, quiz, start with assignments (Wednesdays at 8.45 am) We discuss that week’s material in an interactive setting, there is room for questions, and the lecturer will provide some further clarifying examples. Attending the interactive session only makes sense once you have done the reading and attended the lecture. Once we have gone over the material, we have an individual online quiz to test your understanding of the material. The 10 quiz questions relate to that week’s topic as worked out in the reading/lecture and as discussed at the start of the interactive session. The quiz is not timed; some of you may complete the quiz in five minutes, others in 15 minutes. The quiz Version 220822 is also not graded: its sole purpose is to learn whether you are ready to start working on the assignments. We continue with working on the (group-based) computer assignments of that week, while the lecturer and the TA are present and ready to answer any questions that you may have. 3. Homework (Wednesday-Friday) For the rest of the week, you can work out the computer assignments. Usually, one assignment involves the replication of an article. The assignment includes a couple of question about this article. It will be made clear which pages of the article you have to read. The assignments are due the following week, before the start of Wednesday’s session. Table 1. Schedule of handing in computer assignments Deadline Due Wed Sep 7, 8.45h Computer assignment 1a and 1b Wed Sep 14, 8.45h Computer assignment 2a and 2b Wed Sep 21, 8.45h Computer assignment 3a and 3b Wed Sep 28, 8.45h Computer assignment 4a and 4b Wed Oct 5, 8.45h Computer assignment 5a and 5b Exam The date of the final exam is October TBA, 2022. The exam takes two hours and is either paper-based or is to be taken online. The exam consists of a mixture of open questions, true/false questions, multiple-choice questions and questions that require a numerical answer. In the exam, I present a setting in which someone wants to estimate a treatment effect. Within this setting, among others, you have to show whether you understand what the different research designs can and cannot do for you, how the estimation equation looks like, the assumptions behind each of the designs, differences between ATE and LATE, and how to interpret the estimated coefficients. The exam does not involve any computer-based work with R, but does include questions about how to interpret R output. I will present R output and then ask you to tell me what it means. I am not going to ask you which R command with which options works for a difference-in-difference, for instance. So you should be able to ‘read the language’ not to ‘speak it’. Version 220822 In order to prepare for the exam, you need to read the pages in the textbook and the papers indicated in the syllabus, the slides, the material presented during the lecture that goes beyond what is written on the slides, and you need to study the material from the lab sessions. Make sure to download the latest version of the slides, sometimes changes are made over the course of the semester. The details of the articles you had to read for the assignments are not part of the material. The assignments are meant to expose you to great articles in the field and to understand why authors do what they do. I am not going to ask you about the intricacies of some instrumental variable the authors used in a certain paper, for instance. Grading Your grade is based on your performance on the written exam as well as your participation grade. Your participation grade is based on two parts: 1. The average grade on the graded (group-based) Computer Assignments 1b, 2b, …, 5b. 2. Your grade on pass/fail assignments: the (group-based) Computer Assignment 1a, 2a, …, 5a. You receive a 10 if you complete all elements successfully. For each element that you miss, 2 points will be subtracted from your total, all the way down to a 0. To be able to pass the course, your overall score should be higher than 5.5. Your grade is computed as follows: Grade = 0.75 * Grade Final Exam + 0.25 * Participation Grade Assignment Grade = 0.2 * Average grade a-Assignments + 0.8 * Average grade b-Assignments We have one additional requirement: your score on the final exam should be at least a 5.5. Resit exam If you take the resit exam on January TBA, 2022, your assignment grade still counts. Your grade on the resit should simply be a 5.5 or higher to be able to pass the course. Repeater policy If you are a repeater, then you can choose to either fully redo the course, including all assignments etc., or to do the final exam only. If you go for the first option, then look for other repeaters to make a group of three. Obviously, your participation grade of the previous year no longer counts. Version 220822 Weekly overview Below, we provide an overview of what we do in each week of the course, including any assignments. Week 0. Becoming familiar with R and RStudio Highly recommended for those of you who do not feel confident working with R / RStudio. Absolute must if you do not want to spend an inordinate amount of time on the computer assignments. Please consult the announcement on the Canvas course page for further details. Date: week of Monday, August 22, 2022 Week 1. Selection on observables Session 1: Reading Date: Monday, August 29, 2022 Textbook reading (and links to related videos by the textbook author): Related to the fundamental problem of causal inference: Chapter 5: Identification (https://bit.ly/3pvqr2V, https://bit.ly/3pAfjBW and https://bit.ly/3AC0Co3) Chapter 6: Causal diagrams (https://bit.ly/3T55sBk and https://bit.ly/3Ta9bOm) Chapter 7: Drawing causal diagrams (https://bit.ly/3R2E5q2 and https://bit.ly/3PB6PEZ) Chapter 8: Causal paths and closing back doors (https://bit.ly/3QYdhHd, https://bit.ly/3T6oCGZ and https://bit.ly/3R1WGSK ) Related to functional form: Chapter 13: Regression, as of par. 13.2.2 Polynomials, up to (not including) par. 13.2.4 Interaction terms (https://bit.ly/3dO2Rf5, https://bit.ly/3T7QAlF and https://bit.ly/3pDkz7L) Related to regression – supposed to be familiar material from BSc Econometrics courses: o Chapter 4: Describing relationships (https://bit.ly/3ACYEno, https://bit.ly/3ADTeJ0 and https://bit.ly/3dIGO9x) o Chapter 13: Regression, up to (not including) par. 13.2.2 Polynomials (https://bit.ly/3Cl5Jds, https://bit.ly/3QIgQ4K, https://bit.ly/3QT3Clu and https://bit.ly/3dFvcE9) Other reading (mandatory): Paper that we replicate in Computer Assignment 1b: Dale and Krueger 2002 introduction only, up to page 1494 (link on Canvas). Knowledge clips (if any): see Canvas, module ‘week 1’ Session 2: Lecture Date: Tuesday, August 30, 2022, 08.45-10.30 h Lecture: for slides, see Canvas, module ‘week 1’ Session 3: Computer Assignments Version 220822 Date: Wednesday-Friday, August 31-September 2, 2022 Weekly quiz: Canvas quiz about the material covered this week. Computer Assignment 1a: TBA (assignment available on Canvas). Computer Assignment 1b: work with data on college choices based on Dale and Krueger 2002 (assignment available on Canvas). Week 2. Selection on unobservables: random assignment Session 1: Reading Date: Monday, September 5, 2022 Textbook reading: TBA. Other reading (mandatory): Paper that we replicate in Computer Assignment 2b: Benjamin Edelman, Michael Luca and Dan Svirsky, 2017, Racial Discrimination in the Sharing Economy: Evidence from a Field Experiment, American Economic Journal: Applied Economics, 9 (2), 1-22 (weblink on Canvas). Read up to page 8 only. Knowledge clips (if any): see Canvas, module ‘week 2’ Session 2: Lecture Date: Tuesday, September 6, 2022, 08.45-10.30 h Lecture: for slides, see Canvas, module ‘week 2’ Session 3: Computer Assignments Date: Wednesday-Friday, September 7-9, 2022 Weekly quiz: Canvas quiz about the material covered this week. Computer Assignment 2a: work with data from an experiment among MSc students (assignment available on Canvas). Computer Assignment 2b: work with data from Edelman and Luca (2017) (assignment available on Canvas). Week 3. As-good-as random assignment: instrumental variable approach Session 1: Reading Date: Monday, September 12, 2022 Textbook reading: TBA Other reading (mandatory): Paper that we replicate in Computer Assignment 3a: Daron Acemoglu, Simon Johnson, James A. Robinson, 2001, The colonial origins of comparative development: an empirical investigation, American Economic Review, 91 (5), 1369-1401 (link on Canvas). Knowledge clips (if any): see Canvas, module ‘week 3’ Version 220822 Session 2: Lecture Date: Tuesday, September 13, 2022, 08.45-10.30 h Lecture: for slides, see Canvas, module ‘week 3’ Session 3: Computer Assignments Date: Wednesday-Friday, September 14-16, 2022 Weekly quiz: Canvas quiz about the material covered this week. Computer Assignment 3a: work with data from Daron Acemoglu et al. 2001 (assignment available on Canvas). Computer Assignment 3b: work with data from Joshua Angrist and William Evans, 1998, Children and Their Parents’ Labor Supply. Evidence from Exogenous Variation in Family Size, American Economic Review, 88 (3), 450-477 (assignment available on Canvas). Week 4. As-good-as random assignment: Regression discontinuity design Session 1: Reading Date: Monday, September 19, 2022 Textbook reading: TBA Other reading (mandatory): Paper that we replicate in Computer Assignment 4a: R. Lalive and J. Zweimüller, 2009, How Does Parental Leave Affect Fertility and Return-to-Work? Evidence from Two Natural Experiments, Quarterly Journal of Economics, 124 (3), 13631366 only. (link on Canvas). Knowledge clips (if any): see Canvas, module ‘week 4’ Session 2: Lecture Date: Tuesday, September 20, 2022, 08.45-10.30 h Lecture: for slides, see Canvas, module ‘week 4’ Session 3: Computer Assignments Date: Wednesday-Friday, September 21-23, 2022 Weekly quiz: Canvas quiz about the material covered this week. Computer Assignment 4a: work with data from Lalive and Zweimüller (2009) (assignment available on Canvas). Computer Assignment 4b: work with data from Libertad Gonzalez, 2013, The effect of a universal child benefit on conceptions, abortions, and early maternal labor supply, American Economic Journal: Economic Policy, 5 (3), 160-188. (assignment available on Canvas). Week 5. As-good-as random assignment: Difference-in-differences Session 1: Reading Version 220822 Date: Monday, September 26, 2022 Textbook reading: TBA Other reading (mandatory): Paper that we replicate in Computer Assignment 5a: D. Card and A.B. Krueger, 1994, Minimum wages and employment: a case study of the fast-food industry in New Jersey and Pennsylvania, American Economic Review, 84 (4), 772-793 (link on Canvas). Knowledge clips (if any): see Canvas, module ‘week 5’ Session 2: Lecture Date: Tuesday, September 27, 2022, 08.45-10.30 h Lecture: for slides, see Canvas, module ‘week 5’ Session 3: Computer Assignments Date: Wednesday-Friday, September 28-30, 2022 Weekly quiz: Canvas quiz about the material covered this week. Computer assignment 5a: work with data from Card and Krueger (1994) (assignment available on Canvas). Computer assignment 5b: work with data from Paul Bisschop, Stephen Kastoryano, Bas van der Klaauw, 2017, Street Prostitution Zones and Crime, American Economic Journal: Economic Policy, 9 (4), 28–63 (assignment available on Canvas). ****** Final exam, October TBA, 2022 *****