Course 323063

Fall 2019

Empirical Methods in Finance

Lecturer Part 1: Lecturer Part 2:

Dr. Nicola Pavanini Dr. Julio Crego
Room: K625 Room: K633
Email: [email protected] Email: [email protected]

Teaching Assistants: Lectures’ Time and Rooms:

Ertunc Aydogdu ([email protected]) Tuesday 14:45-16:30, Room GZ 101
Gleb Gertsman ([email protected]) Thursday 10:45-12:30, Room CZ 115
Daniel Karpati ([email protected]) (check schedule for changes!)

Description of the course

This course introduces basic empirical methods that are used in contemporary financial research.
The objective is to provide students with the necessary tools to study the relevant literature of
other core courses in the program and to conduct empirical financial research within the framework
of the Master’s thesis. This is a compulsory course for the Master in Finance. Finance 1 and 2, as
well as some basic mathematical and statistical knowledge are required.
The course pays large attention to the intuition and the applicability of a variety of econometric
techniques that are widely used in contemporary empirical research. Reference will be made to
many real world examples from the corporate finance and asset pricing literature. The classes
intend to provide hands on experience with the econometric package Stata and will focus on a
careful interpretation of the empirical results obtained.

Type of instructions
The course lectures, instructions, and lab hours will be structured as follows:
• Two lectures (HC) per week of 2 hours each taught by one of the lecturers.
• Before the midterm and final exams there will be 2 instructions of 2 hours each, to help
students with the preparation of the exams. The instructions will be taught by one of the
teaching assistants.
• At the end of the first part there will be a STATA mini-course of 1 session of 2 hours, taught
by one of the teaching assistants. During this mini-course students will go over the solution
of a practice assignment together with the teaching assistant. This practice assignment (not
graded) will be posted in advance for students to do it on their own during the computer lab
hours. Towards the end of the second part there will be another STATA based instruction to
help the students with the assignment for the second part of the course.
• 10 sessions of computer lab (LA), of 2 hours each, during which one of the university computer
rooms will be reserved for EMF students to practice with the statistical software STATA and
work on the assignments. These computer lab hours will not have an instructor present.

1
Content
On successful completion of this course, students will be able to:
1. Describe the features of a dataset (e.g. cross section, time series, panel data) and define the
relevant variables.
• This topic requires student to develop the skill to search for and find the appropriate
data that they need to answer a specific research question. It also requires their ability
to perform data manipulation to clean and prepare the dataset for the statistical analysis
they will carry on.
2. Interpret descriptive statistics, such as mean, variance, correlation, and recognize the impor-
tant variables to use.
• This topic requires students to have the ability to describe and analyze the key relevant
information from the dataset they assembled, and especially to highlight the key features
of the data functional to address the research question of interest.
3. Choose the appropriate empirical model (e.g. OLS, Instrumental Variables, Logit, Probit,
Time series or panel estimators) and perform the tests needed (e.g. t-Test, F-Test, het-
eroscedasticity, serial correlation).
• Once students get acquainted with the data, they need to be able to understand what is
the right econometric methodology to use based on the kind of variables in the dataset
and on the research question they are addressing. Moreover, they need to make sure that
the methodology is applied in the correct way, performing all the key statistical tests.
4. Estimate the econometric model using the most convenient estimator(OLS, GLS, etc.).
• Students need to be able to estimate the parameters of the model. Further, they should
be able to perform inference under different assumptions about the error term (hetero-
cedasticity, autocorrelation, cluster correlation, etc.)
5. Analyze the results and compare different models. Comment and interpret the results ob-
tained, based on theoretical hypotheses from the literature and on other empirical studies.
• Once the students obtain an output from the econometric method they applied, they are
required to correctly interpret the results of their model and of the tests they performed.
The interpretation should be both descriptive and critical, as from the results students
should infer any potential drawback of their application.
6. Draw conclusions based on the results obtained to answer the research question.
• Students should give an economic interpretation of their results, being able to provide a
clear answer to their research question, and suggesting any potential policy implication
of their findings.
7. Perform in- and out-of-sample forecast and select the best forecasting model.
• Last, students should be able to deliver forecasts of the explanatory variable and its
variance. Additionally, they need to be able to choose the model that provides the most
accurate forecast according to several criteria (mean squared error, mean absolute error).

2
Materials and Stata
• Lecture notes and papers will be made available via Canvas.
• Main textbook: (W) Wooldridge J.M., “Introductory Econometrics: A Modern Approach”,
(better latest edition); Supplemental material (datasets, exercises) will be posted on Canvas
• Additional textbooks:
– (B) Brooks C., "Introductory Econometrics for Finance", 3rd edition
∗ http://www.cambridge.org/gb/academic/textbooks/introductory-econometrics
for extra resources and Stata guide for all the book’s examples
– (AP) Angrist, J.D., and Pischke J.S., "Mostly Harmless Econometrics"
• Additional references:
– (RW) Roberts, M.R., and Whited T. (2013), “Endogeneity in Empirical Corporate Fi-
nance”, Handbook of the Economics of Finance.
– (BNPW) Bennedsen, M., Nielsen, K., Perez-Gonzalez, F., and Wolfenzon, D. (2007),
“Inside the Family Firm: the Role of Families in Succession Decision and Performance”,
Quarterly Journal of Economics, 122, pp. 647-691.
– (BM) Bertrand, M., and Mullainathan, S. (2003), “Enjoying the Quite Life? Corporate
Governance and Managerial Preferences”, Journal of Political Economy, 111, pp. 1043-
1075.
– (CR) Chava S., and Roberts M.R. (2008), “How Does Financing Impact Investment?
The Role of Debt Covenants”, Journal of Finance, LXIII (5), pp. 2085-2121.
– (MS) Mullainathan S., and Spiess J. (2017), “Machine Learning: An Applied Econometric
Approach”, Journal of Economic Perspectives, 31 (2), pp. 87-106.
• Five Stata web lectures will be posted on Canvas for self-study. The course schedule below
suggests the optimal time to watch them. Students are expected to practice what they learn
in the web lectures during the computer lab hours. These web lectures are:
1. Getting started with Stata
2. Managing and editing datasets
3. OLS regression analysis
4. Panel data regression analysis
5. Time series basics
• Some extra background material to learn how to use Stata:
– Baum C., "An Introduction to Modern Econometrics Using Stata", Stata Press
– http://dss.princeton.edu/online_help/stats_packages/stata/stata.htm
– Machine learning in Stata: https://statalasso.github.io/
• Students have access to a class user account available for the WRDS data library. The sys-
tem can be accessed at: http://wrds-web.wharton.upenn.edu/wrds/index.cfm?true, with the
credentials: Username: emf, Password: WRds2018!

3
Type of Exams
The final course grade will be composed of various parts:
• Midterm Exam: At the end of Part 1 of the course there will be a written midterm exam
covering the content of Part 1. This will be a 1.5 hours exam counting for 30% of the final
course grade. The provisional exam and inspection dates are the following:
– Midterm: 24/10/2019 (Inspection: 6/11/2019)
• Final Exam: At the end of Part 2 of the course there will be a written final exam covering
the content of Part 2. This will be a 1.5 hours exam counting for 30% of the final course
grade. The provisional exam and inspection dates are the following:
– Final: 18/12/2019 (Inspection: 14/01/2020)
• Resit Exam: If the average of the grades of the midterm and final exams is below 50 out
of 100, there will be a written resit written exam counting for 60% of the final course grade
covering the content of both Part 1 and Part 2. The provisional exam and inspection dates
are the following:
– Resit: 21/01/2020 (Inspection: 5/02/2020)
• Assignments: There will be two group assignments, each counting for 20% of the final course
grade, to be done in groups consisting of at least three and at most four students. Groups
must be formed on the Canvas page of the course 2 weeks before the assignment submission
deadline at the latest.1 Assignment grades from past academic years are not valid.
The first assignment will cover the content of Part 1, the second assignment will cover the
content of Part 2. Both assignments are due in digital form on the Canvas page of the course.
Neither late deliveries nor deliveries by email will be accepted. Every student member of
the group must include her/his ANR, name and surname, otherwise the assignment will be
considered as not submitted. You will be given the assignment 2 weeks before the deadline.
The provisional deadlines will be the following:
– First assignment: 15/10/2019 (group formation deadline 30/09/2019)
– Second assignment: 9/12/2019 (group formation deadline 25/11/2019)
Final Course Grade: If the average of the grades of the midterm and final exams is at least 50 out
of 100, the final course grade will be determined for 40% by the grades of the group assignments,
for 30% by the grade of the midterm written exam, and for 30% by the grade of the final written
exam. In order to pass the course, the final course grade needs to be at least 6 out of 10. The
student fails if the average of the grades of the midterm and final exams is below 50 points (out of
100) or if the final course grade is below 6 out of 10. Only the final course grades (the average grade
of the midterm and final written exams, if the student fails to achieve the 50 points mark, or the
average of the midterm and final written exams and the group assignments, if the student achieves
or exceeds the 50 points mark) will be announced by the examination office.
A student that fails needs to take the resit exam. The same rules of the midterm and final written
exams apply to the resit exam, and the grades of the group assignments also count for the resit
exam. In order to pass the course, students need to achieve a minimum of 50 points (out of 100) in
the written resit exam. If a student fails to achieve this mark, the grade will be the one obtained
1
Students and groups that don’t meet these criteria by the group formation deadline will be randomly allocated
to meet these criteria.

4
in the written resit exam. If a student achieves or exceeds this mark, the grade will be determined
as follows: 40% group assignments, 60% written resit exam. Again, in order to pass the course, the
final course grade needs to be at least 6 out of 10.

Part 1: Regression Analysis with Cross Sectional Data

• Week 1:
– 27.08 - Introduction - (W) ch.1
∗ Course organization (material, tutorials, software, assignments, exam)
∗ Introduction to the topics
– 29.08 - Maths and Stats - (W) appendix A-D
∗ Maths and stats foundations
∗ Descriptive statistics

⇒ Watch now Stata web lectures 1,2

• Week 2:
– 03.09 - Bivariate CLRM - (W) ch.2
∗ Bivariate regression model
∗ OLS assumptions and properties
– 05.09 - Bivariate CLRM - (W) ch.3,5
∗ Standard errors
∗ Goodness of fit

⇒ Watch now Stata web lecture 3

• Week 3:
– 10.09 - Bivariate CLRM - (W) ch.4
∗ Hypothesis testing: t-test
– 12.09 - Multivariate CLRM - (W) ch.3,4,
∗ Multivariate regression model
∗ Multivariate OLS assumptions and properties

5
• Week 4:
– 17.09 - Multivariate CLRM - (W) ch.6,7
∗ Hypothesis testing: F-test
∗ Dummy variables
– 19.09 - Causal Inference - (W) ch.8,15; (RW) ch.1,2,3; (AP) ch.3,4; (BNPW)
∗ Instrumental variables

• Week 5:
– 24.09 - Stata Minicourse (Teacher: Daniel Karpati)
∗ Learn how to test the CAPM and the Fama-French 3-factor model in Stata
– 26.09 - Causal Inference - (W) ch.7; (RW) ch.4; (AP) ch.5; (BM)
∗ Difference-in-difference

• Week 6:
– 01.10 - Causal Inference - (RW) ch.5; (AP) ch.6; (CR)
∗ Regression discontinuity design
– 03.10 - Exam Preparation Instruction (Teacher: Ertunc Aydogdu)
∗ Solution of past exams with teaching assistant

• Week 7:
– 08.10 - Discrete Choice Models - (W) ch.7,17
∗ Logit, Probit, Tobit models
– 10.10 - Machine Learning - (MS); https://statalasso.github.io/
∗ LASSO and applications

6
Part 2: Regression Analysis with Time Series and Panel Data

• Week 8:
– 29.10 - Event Study Methodology
∗ Study Design
∗ Abnormal Returns Properties
– 31.10 - Event Study Methodology
∗ Abnormal Returns Properties
∗ Possible Issues in Practice

• Week 9:
– 05.11 - Introduction to Time Series - (W) ch.10
∗ Lags, trends, seasonality
∗ Static and Dynamic models
– 07.11 - Time Series Assumptions - (B) ch.11
∗ Stationarity
∗ Weak Dependence

• Week 10:
– 12.11 - Time Series Processes - (W) ch.18
∗ AR and MA processes
∗ Unit root
– 14.11 - Serial Correlation - (W) ch.17
∗ Serial Correlation Test
∗ Generalized Least Squares (GLS)
∗ Robust Standard Errors

• Week 11:
– 19.11 - Co-integration - (W) ch.18
∗ Unit root
∗ Co-integration
– 21.11 - Co-integration - (W) ch.17
∗ Error correction models
∗ Forecasting

7
⇒ Watch now Stata web lecture 5

• Week 12:
– 26.11 - Volatility Models (B) ch. 8
∗ Heteroscedasticity
∗ Stylized facts
∗ ARCH models
– 28.11 - Volatility Models (B) ch. 8
∗ GARCH models
∗ Conditional heteroscedasticity test
∗ Model selection
– 29.11 - Assignment Preparation Instruction
∗ Solution of past assignments with teaching assistant Daniel Karpati

• Week 13:
– 3.12 - Panel Data Assumptions and Estimators - (W) ch.13 - ch.14
∗ Individual effects
∗ Pooled OLS
∗ First-difference estimator
∗ Within-groups Estimator
– 5.12 - Panel Data Estimators (W) ch.14
∗ Random Effects Estimator
∗ Panel data with groups
∗ Dynamic panel data
– 6.12 - Exam Preparation Instruction
∗ Solution of past exams with teaching assistant Ertunc Aydogdu

⇒ Watch now Stata web lecture 4

• Week 14:
– 10.12 Q&A