Econometrics
II – 2007
Dr.
Liam Delaney
http://www.feweb.vu.nl/econometriclinks/
Please
email me if you have problems downloading
the documents on this webpage.
NEW: Grades from assignments are
available here
NEW: Labs will take place at the
following times (Please Choose One)
Tuesday 10am Daedulus Computer Labs G2
Tuesday 4pm: Daedulus
Computer Labs G1
Course
Lecturer: Liam Delaney (Room B102, UCD Geary Institute, 716-4633. Liam.Delaney@ucd.ie)
Times: Please refer to your Connect
Timetable.
Course Description: This module builds on Econometrics I and introduces some advanced themes in Econometrics. As well as covering theoretical issues in both literatures, students will be made familiar with the practical application of the models using the STATA statistical package. Sample exams are available below and other materials such as further lecture notes, data sets etc will be posted as the course progresses.
Office Hours: I will be available to see students directly after and before the lecture for as long as is required. I will also be contactable by email at Liam.Delaney@ucd.ie and I will be available to see students by appointment if necessary.
Course Objectives: This course aims to introduce the student to advanced econometric topics, both theoretically and practically. The student should, at the end of the course, understand the principles underlying limited dependent variable modelling, simultaneous equation models, panel econometrics and time series econometrics and have applied these principles to economic data sets.
Course Prerequisites: This module builds on Econometrics I and students should be confident that they have a working understanding of hypothesis testing, probability, and in particular the OLS model. Knowledge of calculus is essential. The material for the Econometrics I given by Dr. Vincent Hogan is below.
http://www.ucd.ie/economic/ECON3006/Econometrics1.html
Main
Kennedy, 1998, A Guide to Econometrics, (4th Edition) The
MIT Press,
This book, now in its fourth edition, is a useful guide to the ideas behind the models we are analysing and may be particularly helpful for students who are good verbally but find mathematics difficult.
G.S Maddala “Introduction to Econometrics”, Prentice Hall. This is the main alternative to Gujarati, and students should consult this textbook if they are not satisfied with Gujarati’s discussion of any topic.
Dougherty, C. 2002. “Introduction to
Econometrics” (2nd Edition).
Hill,
R.C.,
Woolridge 2004. “Introductory Econometrics”, Thomson,
South-Western.
Advanced graduate level textbooks that you may wish to consult include:
Verbeek M. 2001. A Guide to Modern Econometrics, Wiley:
Greene, W. 2003. Econometric Analysis, Prentice Hall:
Davidson, R. and
McKinnon, J. 2003. Econometric Theory and
Methods,
Wooldridge, 2002. Econometric Analysis of Cross Section and
Panel Data, MIT Press.
Scott Long, J.
1997. Regression Models for Categorical and Limited Dependent Variables,
SAGE Publications.
Labs: Designated times will be made available to use the computer labs and the course teaching assistants will be present to help you. These labs are vital for you to gain an understanding of how to apply these models, and will be particularly useful exam preparation.
If you need further assistance with STATA, please look here http://www.stata.com/links/resources1.html
Assessment: The course will be assessed on the basis of the annual examination (70 per cent) and the successful completion of four computer assignments (30 per cent).
Please note that the following Lecture notes are provided for your convenience. I will generally (unless otherwise specified) bring a hardcopy for you to the lecture. Some Topics e.g. Topic 5 may take more than one Lecture. These notes are a very poor substitute to actually attending the lectures and I reserve the right to add, subtract, modify etc as the course progresses.
Topic
1 Overview and Review of Classical Linear Regression
This lecture will briefly review what your are expected to know coming in to the course and give a bird’s eye overview of the material that we will cover. Students are expected to know the basics of the classical linear regression model. For this purpose the introductory eight chapters of Gujarati are sufficient. Basic knowledge of dummy independent variables is also assumed. While the level of detail in, for example, Chapter 9 of Gujarati is not explicitly required but you are expected to know how dummy variables (e.g. gender, age categories) are used etc., Part 1 of Wooldridge could also be used to revise the Classical Linear Regression model. Appendix A of Gujarati is good revision as are Appendices A, B and C of Gujarati.
Topic
2 Dummy Dependent Variables (Please email me if you do not have a hard-copy)
Chapter 15 is essential reading. However there are some parts of this chapter that fall outside the scope of our course. The sections on “Grouped or Replicated” Data for both the logit and the probit model can be ignored. The section on “Tobit” models can be ignored for exam purposes but is interesting to read nonetheless. The section on “Modelling Count Data: The Poisson Regression Model” can also be ignored. Section 15.13; “Further topics in Qualitative Response Regression Models” should be read even though you are not required to cover these models in detail. Those of you who want a less mathematical treatment of these topics should consult Dougherty, Chapter 11. Chapter 15 of Kennedy is a neat exposition of general issues with modelling qualitative dependent variables that is good background reading. Kennedy draws the distinction between “Qualitative Dependent Variables” by which he means dependent variables observed in qualitative form e.g. binary and “Limited Dependent Variables” by which he means dependent variables limited in range e.g. attendance data which is limited by full capacity or purchases which have a zero limit. Chapter 17 in Wooldridge is very good and you can use this as a substitute for the chapter in Gujarati if you want to.
All the main textbooks will include empirical applications. Wooldridge is particularly useful for this. A straightforward example of the probit and logit model in practice is:
Tillman, R., and Pontell, H., (1995). "Organizations and Fraud in the Saving and Loan Industry". Social Forces, Vol. 73. No. 4. 1439 - 1463 (available from JSTOR)
For those of you who want a more advanced treatment, the standard textbook is Greene, Econometric Analysis. For those of you considering postgraduate work in Economics, this book is particularly useful. The first four chapters of Scott-Long (1997) are also very good and give a simplified overview of how maximum likelihood is actually computed for those who are interested. Other advanced treatments are listed below. These are completely optional readings.
#1. "Discrete Choice
Methods with Simulation" by Kenneth E. Train.
#2. Econometric Analysis of Discrete Choice
Publisher: Springer Verlag (December, 1987)
ISBN: 3540185348
#3. Applied Choice Analysis :
A Primer
by David A. Hensher,
#4. Economic Choice Theory :
An Experimental Analysis of Animal Behavior by
Publisher:
ISBN: 0521454883
#5. Models for Discrete Data
Publisher:
ISBN: 0198524366
#6. Categorical Data Analysis (Wiley Series in
Probability and Statistics)
by Alan Agresti
#7. An Introduction to Categorical Data Analysis
by Alan Agresti
Topic 3
Simultaneous Equations
Chapters 18, 19 and 20 of Gujarati are essential reading. All of Chapter 18 is required knowledge and you should familiarise yourself with each of the examples to gain an understanding of the variety of circumstances under which simultaneity can arise. Once again, all of Chapter 19 is essential. We did not cover the Hausman test or the Rank condition for identification in lectures and you are not required to know about these in great detail for the purpose of this module but you are advised to read through these topics to understand the core material in more depth. All of Chapter 20 is essential reading. Chapter 14 of Hill, Griffiths and Judge (2001) is worth reading. This contains an explanation of the truffles example that we used in the tutorial. Chapter 10 of Kennedy (4th Edition) is highly recommended. His discussion of Identification is excellent and essential reading. For those seeking a more basic introduction, Dougherty’s chapter on simultaneous equation modelling is advised. Chapter 16 of Wooldridge is also very good. Many of the examples we examined in class come from this treatment. Those looking for a more advanced treatment should consult Greene. The paper by Romer (1993) that we talked about in class is:
Romer, D., (1993), "Openness and Inflation", Quarterly Journal of Economics, 108, 869 - 903.
The Levitt paper is:
Both papers are available (via JSTOR) at http://www.jstor.org/journals/00335533.html
Topic 4
Distributed Lag Modelling
Chapter 17 of Gujarati is essential reading for this topic. Gujarati does not discuss arithmetic lags. The lecture notes are sufficient for arithmetic lag structures. This entire chapter is essential reading. The material on the Koyck and Almon models is particularly good. The Almon material is the equivalent of the discussion in the lecture notes of polynomial distributed lag models. The discussion of partial adjustment and adaptive expectations models is essential. You are advised to read the sections on Granger causality and exogeneity to further your understanding, although detailed knowledge of these topics is beyond the scope of this module. The same applies to the SARGAN test. Chapter 17 of Wooldridge and Chapter 15 of Hall, Griffiths and Judge provide an alternative treatment to that of Gujarati. Patterson’s book has a number of excellent chapters on empirical examples that are great background reading. His chapter on the demand for money is excellent
Topic
5 Time Series
Chapters 21 of Gujarati is essential reading for these topics. Chapter 22 is helpful for certain topics particularly VAR analysis, but as discussed in lectures we will not be covering ARCH, GARCH models. While the material is ordered differently in the textbook almost all of the content of Chapter 21 is covered in the lectures. An excellent book for Time Series that I would recommend is Kerry Patterson’s “An Introduction to Applied Econometrics: A Time Series Approach” which discusses all of the topics outlined in detail in a relatively non-intimidating fashion. Chapter 17 in Kennedy (4th Edition) is worth reading for background purposes. Chapter 12 and Chapter 14 of Dougherty’s text are also good introductions. Ender’s “Applied Econometric Time Series” (2nd Edition) is a graduate level textbook which will suit those who want to progress on to more technical models.
Topic 6
Introduction to Panel Econometrics
Chapter 16 of Gujarati (4th Edition). The chapter is short and relatively straightforward and a careful and close reading of its content will provide a good introduction to the very large and complex area of panel econometrics. Those looking for an advanced treatment should consult Greene. Baltagi’s Econometric Analysis of Panel Data is also highly recommended. The empirical examples used in Wooldridge and Green are ample to provide discussion in your essays.
Topic
7 Review
Assignments
The assignments will be announced in class. For each assignment, I will place a copy of the assignment, the data-set and the date on which the work is due in, in the space below. Assessment assignments will be made available on the day of the assessment at the appointed class time.
|
Assignment |
Data |
Assignment Date |
|
1. Limited Dependent Econometrics b. Assessment Assignment 1a c. Assessement Assignment 1 b |
Bio-Chemists Data Labour Force Participation Data |
|
|
2. Instrumental Variables |
Cigarette data (same data as above) |
|
|
3. Distributed Labs c. Assessment Assignment 3b |
|
|
|
4. Time Series (Practice Only) |
|
|
|
a. Practice Assignment |
|
|
Last years exam and two sample exams are provided below.
Nobel Laureates in Econometrics
Jan Tinbergen and Ragnar Frisch were awarded in 1969 (the first Nobel Price for Economic Sciences) for having developed and applied dynamic models for the analysis of economic processes
Lawrence Klein was awarded in 1980 for his computer modelling work in the field.
Daniel McFadden and James Heckman shared the award in 2000 for their work in microeconometrics.
Robert Engle and Clive Granger were awarded in 2003 for work on analysing economic time series. Engle pioneered the method of autoregressive conditional heteroskedasticity (ARCH) and Granger the method of cointegration.
Finn Kydland and Edward Prescott were awarded the prize in 2004 for their contributions to dynamic macroeconomics: the time consistency of economic policy and the driving forces behind business cycles