Economics 506 - Mathematicial Statistics - Professor Norman R. Swanson

Note: I have not taught this course in many years.

Econ506 is the first course in the core econometrics sequence. The main purpose of this course is to provide you with a rigorous foundation in statistics and probability, enabling you to successfully complete the rest of the core in econometrics. Focus in the course will be on basic principles, including among other things: probability, random variables, conditional probability, probability densities and distributions, characteristic functions, test statistic formulation and distribution theory, statistical inference, and basic regression. Time permitting, we will also cover basics in Monte Carlo experimentation, bootstrapping, and nonparametric testing. Although a firm theoretical foundation in statistics is viewed as essential to continuing with the core, it is assumed that many of you will ultimately be interested in pursuing applied fields in economics. For this reason, some emphasis will be placed on applied problem solving using the tools learned in the class. In this regard, I will teach basics of computing using the GAUSS computer program. GAUSS will be used for assignments in Econ506, as well as in later core econometrics courses.


·  Monday October 28: In class test I -- approx. 1:20 in length

·  Monday December 2: In class test II -- approx. 1:20 in length

·  Monday October 14: Assignment 1 outlined in class.

·  Monday October 21: Assignment 1 is due in class.

·  Monday November 18: Assignment 2 outlined in class.

·  Monday November 25: Assignment 2 (i.e. the final project) is due on the last day of classes (Monday December 9).

·  Monday December 9: Last Class!!


·  Course Syllabus

·  Economics 506 Course Syllabus, Fall 2002. PDF Version.

·  Practice and Text Problems and Solutions

·  Economics 506 Practice Problems I. PDF Version.

·  Economics 506 Practice Problems I - solutions. Document.

·  Economics 506 Practice Problems II. PDF Version.

·  Economics 506 Practice Problems II - solutions. Document.

·  Economics 506 Text Practice Problem Solutions - Sections 3 and 4 of Syllabus. Document.

·  Economics 506 Text Practice Problem Solutions - Section 5 of Syllabus. Document.


·  Gauss -- in class demo (Monday October 7)

·  Eviews - in class demo (Monday October 28)


·  Gauss Monte Carlo Experiment Example

·  Gauss Program #1

·  Gauss Program #2

·  Gauss Program #3


·  Mark Watson's Gauss Tutorial

·  Eric Zivot's Gauss Resources

·  Paul Söderlind's Software Page

·  Gary King's Gauss Programs

·  Index of Gauss Archives

GRADES FOR ECMT506 (when available)

·  Grades by last 4 digits of SSN

ASSIGNMENTS: (These are the assignments for Fall 2002)


I would like each of you to collect an economic dataset from any source of your choosing (there is a link at the end of this page to many data links). Your dataset should have at least 3-4 variables in it, and each variable should have at least 100 observations. I would like you to examine your data and do the following: 1. Graph the data and discuss economic episodes in the data, apparent structural breaks, possible outliers, etc. 2. Construct basic statistics for all series, including mean, variance, covariances, and correlations. Discuss these, and construct confidence intervals when possible, etc. 3. Given your findings in 1 and 2 above, hypothesize, both based on economic intuition and based on your statistical calculations what the relationship(s) between the variables might be. Discuss the possibility of fitting linear and nonlinear models, and then proceed using least squares estimation to fit at least 3 different models to your data, choosing a ``target variable" that you wish to explain via use of the other variables (and/or lags of the ``target" and other variables). Discuss the results of your regressions via examination of t- and F-tests, as well as correlations, etc.

Particulars: I do not wish to receive from you great collections of computer output. Rather, please interpret your own computer output, and hand in a summary of your results and findings to me, including garphs of the data, tables of statistical output, etc.


I would like you to carry out a Monte Carlo assessment of the finite sample properties of a statistical test. This should inlcude examination of the empirical level and power of a test for various sample sizes, model parameterizations, and nominal test sizes. You should form a basic parameter set with which to calibrate your data generating processes via empirical (regression) analysis of your dataset from Assignment 1. Additionally, the test that you examine should be one that has been of use to you in your analysis of your dataset, and you should finalize with me the details concerning which test you wish to examine. For those of you who want to do something more advanced, you may instead compare the performance of a bootstrap procedure for calculating test critical values with tests constructed using critical values derived from the usual asymptotic approximations. Again, please discuss with me first.

Particulars: Monte Carlo results should be presented in tabulated form, along with detailed explanation of which models were fitted, what parameterizations were used when generating data, the properties (asymptotic) of the test being examined, etc.


Assignment 2 serves as the final project for this class.


·  Summary of Various Discrete and Continuous Distributions.

·  Gauss example program - Monte Carlo, Basic Regression, and Statistic Calculation.

·  Eviews overview.

·  papers as they are discussed in class


·  Data Links, Econ Dept's, Conferences, and Economists


·  Any queries or comments? Please click here ---> ---

·  Back to my home page