Econometrics (MSc) Kurt Schmidheiny Introduction 2

Fall 2022 University of Basel

Your Teacher
Prof. Dr. Kurt Schmidheiny
Universität Basel
Peter Merian-Weg 6, Office 5.55
kurt.schmidheiny(at)unibas.ch
Econometrics (MSc) Office Hours: Tuesday afternoon by appointment per email

Teaching Assistant:
Andrea Ghisletta
Universität Basel
Peter Merian-Weg 6, Office 5.49
andrea.ghisletta(at)unibas.ch

Version 20-9-2022, 21:25

Introduction 3 Introduction 4

Class Schedue Course Homepage

Lecture: There is a course homepage with slides, handouts and additional readings:
Mon 10:15 - 12:00 (Biozentrum, U1.131)
http://www.schmidheiny.name/teaching/unibas/econometrics/
Wed 10:15 - 12:00 (Biozentrum, U1.131)
(username: unibas; password: )
The course is planned as an in-person course.

Two extra Lectures:

Wed 09/11/2022 12:15-14:00 (Biozentrum, U1.141)
Wed 07/12/2022 12:15-14:00 (Biozentrum, U1.141)

Exercise sessions:
During lecture time. Will be announced as we go ahead.
Introduction 5 Introduction 6

Introduction Challenge of this course: Heterogeneity

Most widely used econometric tools: This is a very large class with students from very different backgrounds and
OLS, IV, 2SLS, FE, RE, Probit with very different goals.

• What is it?
• How are they used?
• When can they be used?
• When can they not be used?
• What can go wrong?

Introduction 7 Introduction 8

Heterogenous Backgrounds Heterogenous Goals

Some of you ... Some of you ...

... have taken several courses in econometrics ... want to do empirical research in their PhD
... have only taken basic statistical courses ... want to a purely theoretical PhD

... very much like mathematics ... want to use quantitative analysis in their work,
e.g. consultants, traders, financial analysts
... very much like getting results
... want to become deciders
... have performed sophisticated own empirical projects
→ Some disappointment is inevitable.
... have never (really) run a regression
⇒ I want to challenge all of you in some dimension.
⇒ You will be able to (partly) choose the formal level.
Introduction 9 Introduction 10

Outline Outline
1. Causal effects and the logic of randomized experiments + Data generating process and Monte Carlo
2. Linear regression:
+ Sampling distribution and asymptotic properties
Estimation, small and large sample properties, hypothesis testing, omitted
variable bias, model selection, functional form + Maximum likelihood estimation
3. Robust inference in the linear model:
heteroscedasticity and clustering

4. Instrumental variable estimation:

Estimation, identification, weak instruments

5. Panel data: Fixed effects, random effects

6. Binary choice: probit and logit

Introduction 11 Introduction 12

Companion Course Companion Course (cont.)

Advanced Econometrics 4. The algebra of the multivariate linear regression: degrees of freedom,
Wednesday, 14:30-16:00 (WWZ S16/S14) Gauss-Markov theorem, Frisch-Waugh-Lovell theorem
by Prof. Schmidheiny
5. The algebra of instrumental variable estimation
1. Elements of matrix algebra: basic operations, trace, rank, inverse, 6. The algebra of basic panel data methods: within and between
eigenvalue and spectral decomposition transformation, testing for unrelated effects under non-spherical
2. Elements of probability theory: random variables, joint, conditional and disturbances
marginal distribution, expected value and other moments, change of 7. Maximum Likelihood Estimation
variables
8. Binary choice as an example of deriving estimators and their properties
3. Elements of statistics: point estimation, interval estimation, hypothesis
using maximum likelihood
testing, large sample theory
Introduction 13 Introduction 14

This course deals (mainly) with observational data Level of the course
This course deals with Data which is non-experimental, i.e. not from The level of this course is between introductory and advanced textbooks.
experiments
It is introductory concerning ...
• Data from surveys, public records, accounting, ... • ... most of its topics
• Traditional approach of econometrics • ... its mathematical rigor (limited use of matrices)
• Prevailing in empirical literature
It is advanced concerning ...
• Widely used in current research
• ... some of its topics
But experiments become more and more important.
• ... its mathematical rigor (we do some proofs)
• ... the applications

Introduction 15 Introduction 16

Level of the course (cont.) Prerequisites

“Econometrics” I assume that you took an introductory course in statistics/econometrics.
+ “Advanced Econometrics”
In particular, you need to know basics of probability
= Really “Advanced Econometrics”
• discrete and continuous random variables
• probability density function, cumulative distribution function
• expected value, E(.), variance, V(.), covariance
• rules about them
⇒ Read Stock/Watson, section 2.1-2.4
Introduction 17 Introduction 18

And basics of statistics Introductory textbooks

• estimation of mean I Stock, James H. and Mark W. Watson (2020)
• standard error, confidence intervals Introduction to Econometrics, 4th Global ed.
Pearson
• hypothesis testing, t-test, p-value
I Wooldridge, Jeffrey M. (2009)
⇒ Read Stock/Watson, section 3.1-3.7
Introductory Econometrics: A Modern Approach, 4th ed.
Cengage Learning
I Wooldridge, Jeffrey M. (2014)
Introduction to Econometrics, 5th ed.
Cengage Learning EMEA
(Identical content as 2009 but no appendices)

Introduction 19 Introduction 20

Advanced textbooks Companion textbooks

I Cameron, A. Colin and Pravin K. Trivedi (2005) I Angrist, Joshua D. and Jörn-Steffen Pischke (2009)
Microeconometrics: Methods and Applications Mostly Harmless Econometrics: An Empiricist’s Companion
Cambridge University Press Princeton University Press
I Wooldridge, Jeffrey M. (2002) I Kennedy, Peter (2008)
Econometric Analysis of Cross Section and Panel Data A Guide to Econometrics, 6th ed.
MIT Press Blackwell Publishing
Introduction 21 Introduction 22

Handouts Matrices or NO matrices?

There are handouts for all topics of the course. How should you choose the formal level of this course?

These handouts are ... • You can pass the exam with or without matrices. You may not reach to
... very brief maximum number of points. But almost.
... not self-contained ⇒ Choose your level based on your background and/or ambition
... intended to be a useful companion for your life after this course • Only choose matrices if you feel comfortable with them. Spend your time
⇒ You will absolutely need to work with one or more textbooks on the econometrics and not on the mathematics.
... most handouts will come in two versions:
with use of matrices and without
• Choose matrix-free if you just want to pass this exam.
• Choose matrices if you want to continue with Time Series Analysis I/II or
Microeconometrics I/II.
.

Short Guides to Microeconometrics Kurt Schmidheiny

Fall 2022 Unversität Basel

The Multiple Linear Regression Model

Introduction 23 Introduction 24

1 Introduction
Statistical Software Problem Sets
The multiple linear regression model and its estimation using ordinary
least squares (OLS) is doubtless the most widely used tool in econometrics.
• We will
It use
allows R
to estimate the relation between a dependent variable and a set
There will be 8 problem sets.
of explanatory variables. Prototypical examples in econometrics are:
• Handouts provide R and STATA code • They will not be graded
• Wage of an employee as a function of her education and her work • You don’t have to hand them in
• We assume you
experience (theare familiar
so-called Mincer with some statistical software
equation).
• They will be discussed by the teaching assistant in dedicated lectures
• Price of a house as a function of its number of bedrooms and its age
Alternative: (an example of hedonic price regressions). during the usual time slots.
The dependent variable is an interval variable, i.e. its values represent
• Usea another statistical package.
natural order and differences of two values are meaningful. In practice, Note:
this means that
For example, STATA the variable needs to be observed with some precision
and that all observed values are far from ranges which are theoretically • It is in your own interest to do the problem sets seriously
• excluded.
Please checkWages,
withfor example,
me if itdocovers
strictly speaking not qualify we
all methods as they
use. • The problem sets are key for your understanding
cannot take values beyond two digits (cents) and values which are nega-
tive. In practice, monthly wages in dollars in a sample of full time workers • They are very important for the exam
is perfectly fine with OLS whereas wages measured in three wage cate-
gories (low, middle, high) for a sample that includes unemployed (with • Doing the problem sets in groups can considerably increase their value
zero wages) ask for other estimation tools.
added
Introduction 25 Introduction 26

Online Tests Exam

There will be an online test for each problem set. There will be a final exam in January
• They have to be completed before the discussion in class
• Part with Multiple Choice questions
• They will be graded on a pass / fail basis
• Part with open question based on output from statistical software
• You must pass at least 5 out of the 8 online tests in order to be allowed to
the final exam • No proofs.

Note:
• The questions in the online test are different from the questions in the
problem set.
• The online test is a good preparation for the problem set.
• The questions in the online tests are generally easier than the exam
questions.