Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit

Economics 140 – Fall 2021

Course Syllabus

Welcome to Economics 140! This course is meant to introduce you to the statistical analysis of economic data, also known as “Econometrics.” You should have taken both microeco-nomics and macroeconomics (either the Economics 100A/B or 101A/B series, or equivalent). More importantly, you have completed an introductory statistics course, e.g., Statistics 2, 20, W21, or an equivalent. On occasion, when it helps to explain a concept, we will make use of calculus – which is not a pre-requisite for the course, but it is a requirement for the Economics major. Those looking for a more mathematical and theoretical treatment of the same topics covered in this course are encouraged to take Economics 141.

After reviewing essential concepts from probability and statistics, we turn to the heart of the course: regression analysis. You will learn not only the meaning and properties of both univariate and multivariate regression, but also how to test economic relationships using real datasets and an econometrics software package. We will develop techniques to handle common statistical problems that arise when working with economic data including endogeneity, selection bias, mis-specification, and measurement errors. We will expand the types of data we can analyze by exploring the topics of panel data, binary response, and quasi-experiments.

General Information

● Instructor: Stephen Bianchi, 673 Evans & Zoom

● Class Meeting Times: TTh 2-3:30pm, Stanley 105 & Zoom

● Office Hours: W 3:30-5pm, 673 Evans & Zoom

● Enrollment: Please see the Economics Department Head GSI, John Wieselthier (548 Evans, [email protected]), for ALL questions regarding enrollment.

● Email: [email protected]

● Email Policy: When you email me, please put “[ECON140]” in the subject and ask me questions that can be answered in a few sentences. If I find that my response will require more than a few sentences, I will ask you to come see me during office hours. I will reply to course related emails within 48 hours.

● Discussion Sections: Due to remote instruction, you need NOT attend your first section meeting in order to remain enrolled in the course. This is a change from past semesters with on-campus instruction. Each GSI is only responsible for students who are officially registered in one of their sections, so please do not email another GSI. However, you may go to any GSI’s office hours. If you have a conflict, you may also attend your GSI’s other regularly scheduled section – but before doing so, please discuss with your GSI.

● Accomodations: If you need disability-related accommodations in this class, if you have emergency medical information you wish to share with us, or if you need special arrangements in case the building must be evacuated, please inform John Wieselth-ier immediately. For disability-related accommodations, you must also obtain a Let-ter of Accommodation (LOA) from the Disabled Students’ Program (http://dsp. berkeley.edu), which they send electronically to me. Request for exam accommoda-tion must be received and acknowledged by me or Alexey at least two weeks before an exam, which is DSP’s own internal deadline for scheduling the proctoring of exams. Accommodations are not offered retroactively.

● Academic Honesty: In fairness to students who put in an honest effort, cheaters will be harshly treated. Any evidence of cheating will result in a score of zero on that assignment. Cheating on the midterm or the final exam results in an “F” for the course. Cheating includes but is not limited to bringing unauthorized written or electronic materials into an exam, using unauthorized written or electronic ma-terials during an exam, copying off another person’s exam or assignment, allowing someone to copy off of your exam or assignment, having someone take an exam or assignment for you, changing an exam answer after an exam is graded, and plagia-rizing written or other materials. Incidences of cheating are reported to the Cen-ter for Student Conduct, which administers additional punishment. See also http://sa.berkeley.edu/conduct/students/standards.

● Limits to Confidentiality: As UC employees, all course instructors and tutors are Responsible Employees, and we are required to report incidents of sexual violence, sexual harassment or other conduct prohibited by university policy to the Title IX officer. We cannot keep reports of sexual harassment or sexual violence confiden-tial, but the Title IX officer will consider requests for confidentiality. There are confidential resources available to you, including the CARE Advocate Office (http://sa.berkeley.edu/dean/confidential-care-advocate), which serves survivors of sexual violence and sexual harassment.

● Honor Code: We at UC Berkeley have adopted this Honor Code: “As a member of the UC Berkeley community, I act with honesty, integrity, and respect for others.” Your Econ 140 instructors join you in pledging to adhere to this code.

Course Books

Required:

● (MM) Joshua D. Angrist and J¨orn-Steffen Pischke, Mastering ’Metrics: The Path From Cause to Effect, 1st Edition.

Optional (but strongly recommended):

● (SW) James H. Stock and Mark W. Watson, Introduction to Econometrics, 3rd or 4th Edition.

There is a companion website for the 4th edition of Stock & Watson at: https://www.princeton.edu/~mwatson/Stock-Watson_4E/Stock-Watson-Resources-4e.html. Many study resources are available on this site including answers to end-of-chapter questions, datasets for empirical exercises, replication files for empirical analyses reported in the text-book, and additional empirical exercises. There is also a companion website for the 3rd Edi-tion at: https://wps.pearsoned.com/aw_stock_ie_3/178/45691/11696965.cw/index.html.

Course Software

The assignments in this course will be in Jupyter Notebooks using Python. Python is a general purpose open-source programming language utilized commonly by economists, data scientists, and programmers alike. We will primarily use the statsmodels library to carry out econometric analyses, in addition to pandas for data manipulation. All assignments will be distributed and completed in Jupyter Notebooks, an intuitive and interactive computing environment that contains both text and code.

Your notebooks will be hosted on DataHub, a free campus-wide cloud service that will provide the computing environment for your code. This means that you don’t have to install anything on your computer; instead you can access all assignments through a browser (preferably Google Chrome). If you have taken Data 8 or other data science courses on campus, the format should be familiar to you.

Do not worry if you have never used Python before; the first (optional, but highly rec-ommended) assignment in the class will help familiarize you with Python and the Jupyter environment. There are many other full-service econometrics packages (e.g., Stata, R, Mat-lab, SAS) but these will not be supported by your GSIs. Experience with Python can be helpful if you do other economic research (e.g., an honors thesis) and it looks good on your job resume.

Requirements

The course requirements include one midterm, a final, and five graded problem sets. The course grade will be determined as follows:

● Problem Sets (40%)

● Midterm (25%)

● Final (35%)

Problem sets: You are encouraged (but not required) to form study groups of up to three students. The group may submit a single answer sheet with the names of all of the study group members at the top of the first page. Everyone in the study group receives the same grade. We will use the usual 3-point “check” system of grading problem sets.

Answers to problem sets must be submitted via Gradescope by the specified time on the due date. No late work will be graded and, yes, that penalizes all members of the study group. Problem set 1 will be posted after the first lecture and is due Tuesday, September 14th.

Exams: There will be a midterm exam on Thursday, October 21st and a final exam on Monday, December 13th. If you do relatively better (i.e., earn a higher standardized score) on the final than on the midterm, your final score will count for 60% of your overall class score.

Dates for exams will not change and make-up exams will not be given. If you fail to take the midterm (for any reason), your final exam will count for 60% of your overall class score. If you fail to take the final (again, for any reason), you must petition for an incomplete. But please note that incompletes will not be granted unless you meet the University standards and those have become increasingly demanding. If you do not take the final and do not petition for an incomplete you will receive an “NP” or an “F” for the course (depending on your grading option).

Course Outline

The following is a tentative schedule of the topics to be covered in this class – it is likely to change a fair amount as we progress. The corresponding readings are from Angrist & Pischke (MM) and Stock & Watson, 4th Edition (SW). Lectures will loosely follow these readings.

● Classical Statistics & Simple Regression

– Lecture 1 (August 26): introduction, data types, random variables

Readings: (MM) Introduction, Chapter 1; (SW) Chapter 1, Sections 2.1, 2.2

– Lecture 2 (August 31): random variables, probability distributions, random sampling and sample average

Readings: (MM) Chapter 1; (SW) Sections 2.3, 2.4 (pp. 33-35), 2.5, Appendix 2.1

– Lecture 3 (September 2): convergence of random variables, law of large num-bers, central limit theorem, hypothesis tests

Readings: (MM) Chapter 1; (SW) Sections 2.6, 3.1

– Lecture 4 (September 7): t-tests, p-values, confidence intervals, testing for difference in means, multiple random variables, joint probability distributions, conditional probability

Readings: (MM) Chapter 1; (SW) Section 3.2, 3.3, 3.4, 3.7, Appendix 3.2

– Lecture 5 (September 9): interpreting statistical evidence, conditional expec-tation, law of iterated expectations (LIE), economic relationships and the condi-tional expectation function (CEF)

Readings: (MM) Chapter 2 (Appendix); (SW) Section 3.5

– Lecture 6 (September 14): CEF decomposition property, CEF prediction property, bivariate linear regression and the CEF, bivariate linear regression (es-timation), problem set 1 due

Readings: (MM) Chapter 2 (Appendix); (SW) Section 4.1, 4.2

– Lecture 7 (September 16): linear regression model, unbiasedness and asymp-totic normality of OLSEs, goodness of fit 

Readings: (MM) Chapter 2; (SW) Section 4.3, 4.4, 4.5

– Lecture 8 (September 21): goodness of fit, regression with binary independent variable and relation to difference in means testing

Readings: (MM) Chapter 2; (SW) Sections 5.1, 5.2, 5.3

– Lecture 9 (September 23): regression with binary independent variable, one-sided hypothesis tests, heteroskedasticity and homoskedasticity, nonlineariteis and simple linear regression

Readings: (MM) Chapter 2; (SW) Section 5.4

● Multivariate Regression

– Lecture 10 (September 28): Gauss-Markov Theorem, multiple linear regres-sion, Frisch-Waugh Theorem

Readings: (MM) Chapter 2; (SW) Sections 5.4, 6,1, 6.2, Appendix 6.3

– Lecture 11 (September 30): comparisons of univariate and multivariate coef-ficent estimates, multicollinearity

Readings: (MM) Chapter 2; (SW) Sections 5.4, 6,1, 6.2

– Lecture 12 (October 5): irrelevant variables, omitted variables, goodness of fit in multiple linear regression, problem set 2 due

Readings: (MM) Chapter 2; (SW) Sections 6.3, 6.4, 6.5, 6.6, 6.7

– Lecture 13 (October 7): multivariate regression: Gauss-Markov Theorem, hypothesis testing

Readings: (MM) Chapter 2; (SW) Sections 7.1, 7.2, 7.3

– Lecture 14 (October 12): multivariate regression: hypothesis testing

Readings: (MM) Chapter 2; (SW) Sections 7.1, 7.2, 7.3

– Lecture 15 (October 14): specifying a multivariate regression model, multi-variate regression: incorporating nonlinearities

Readings: (MM) Chapter 2; (SW) Section 7.5, 8.1, 8.2

– Midterm Review (October 19): catch up and review, problem set 3 due

– Midterm (October 21): 2-3:30pm (Pacific Time), no lecture

– Lecture 16 (October 26): multivariate regression: single and mutliple binary independent variables, interaction terms

Readings: (MM) Chapter 2; (SW) Sections 8.2, 8.3

– Lecture 17 (October 28): internal and external validity

Readings: (MM) Chapter 2; (SW) Sections 9.1, 9.2

● Fixed Effects, Binary Response, Instrumental Variables

– Lecture 18 (November 2): panel data, entity fixed effects

Readings: (SW) Sections 10.1, 10.2, 10.3, 10.4, 10.5

– Lecture 19 (November 4): panel data, time fixed effects

Readings: (SW) Sections 10.1, 10.2, 10.3, 10.4, 10.5

– Lecture 20 (November 9): limited dependent variable models, linear proba-bility model

Readings: (SW) Section 11.1

– Veterans Day (November 11): no lecture

– Lecture 21 (November 16): logit and probit models, binary dependent variable models: goodness of fit, estimating partial effects

Readings: (SW) Sections 11.2, 11.3

– Lecture 22 (November 18): instrumental variables and two stage least squares, problem set 4 due

Readings: (SW) Section 12.1

– Lecture 23 (November 23): instrumental variables and two stage least squares

Readings: (SW) Section 12.2

– Lecture 24 (November 30): efficiency of OLS estimators versus IV estimators, testing overidentifying restrictions

Readings: (SW) Section 12.3

– Lecture 25 (December 2): potential outcomes, causal effects, differences-in-differences estimator

Readings: (SW) Sections 13.1, 13.2, 13.3, 13.4

– Final Review (December 7): catch up and review, problem set 5 due

● Final Exam (December 13): 3-6pm (Pacific Time)