Course: ECO302
Instructor: Kirill Evdokimov
F F2017

### Description of Course Goals and Curriculum

ECO 302 is one of the three required core departmentals for Economics concentrators. 302 is designated as the less mathematically rigorous version of econometrics in comparison to its more mathematical counterpart, ECO 312. The class aims to help students develop a facility with basic econometric methods and the ability to apply them to actual problems and understand their application in other substantive course work in economics. The textbook used for this class is Stock and WatsonIntroduction to Econometrics (3rd Edition), and the course covers chapters 1 - 12, encompassing topics such as:
• Linear regression, hypothesis tests
• BLUE estimators
• Multiple regressions, joint hypotheses and F-tests
• Non-linear regressions, polynomials, interaction terms
• Probit and logit regressions
• Panel data, time and entity fixed effects, difference-in-difference
• Regression discontinuity design
• Endogeneity issues
• Instrumental variables
The first half of the course builds on materials explored in ECO 202 (such as estimating the slope coefficient and finding confidence intervals). However, the focus of 302 is more theoretical in that it aims to show students the motivations behind the construction of certain parameters and methods of regression and demonstrating why they are unbiased. Students should not expect a plug-and-chug statistics class that involves finding the right command in the calculator and simply regurgitating identities shown in class. Apart from econometric proofs, a substantial practical component of this class involves the use of STATA (a data analysis package) which are used heavily on problem sets. Students may be asked to reproduce certain commands to achieve a desired regression on the final, although STATA are not usually heavily tested in exams.

### Learning From Classroom Instruction

Slides used in class are often a condensed version of the text found in the textbook, often using the same numerical examples. It is helpful to have read the chapter before coming to lecture, though this is not essential. A recurring theme in this course is demonstrating/proving why a certain parameter/estimator is unbiased using the 3 or 5 least-squared assumptions such as conditional mean of u being 0, X and Y being identically and independently distributed, finite fourth moments, and homoskedasticity. This is a skill that need to be practiced, especially in novel situations where some assumptions may not hold. As the class goes on, it is helpful to keep a record of all the identities and formulae used in proofs. As preparation for the midterm and final, students are allowed to bring a "cheat sheet" (or 2 in the latter) of formula they deem necessary.

### Learning For and From Assignments

Problem sets are a mix of even-numbered questions from the books, as well as others written by the instructor himself. It is likely that the difficulty of the exams will be comparable to the latter and not the former. As a result, it is helpful to do the practice midterm and final as part of the preparation.

### External Resources

Since most of the econometric content covered in this class is still quite introductory, a plethora of other online sources are available to aid students' understanding.

### What Students Should Know About This Course For Purposes Of Course Selection

If students are taking this class, it is likely because they have to fulfill the core Economics department. Those interested in more sophisticated and advanced econometrics may benefit from taking other more rigorous classes such as ECO 312 and those in the ORF department. Those interested in 312 may find it helpful to obtain a copy of both syllabi to compare the differences between the two courses, namely:
• 312 uses linear algebra/matrix notation for regressions with many regressors. Properties such as the conditional of mean of u being 0 are expressed correspondingly with linear algrebraic properties such as orthogonality. While taking MAT 202/204 is recommended, the class is designed so that students who are willing to make the jump can do so within reasonable effort and time
• 312 covers extra topics that are not covered in 302, such as quasi-experiments and time series. As such, the pace with which 312 progresses is faster than that of 302, so this should be kept in mind while designing your course load for the semester.
• Those wishing to take ECO 313 may find it more beneficial to enroll in 312, since the continuation between these two classes will be more seamless. ECO 313 completes the remaining chapters of Stock and Watson.
Econometrics