Centre for Financial & Management Studies (CeFiMS) - University of London

Individual Professional Courses – IPC  

Econometric Principles & Data Analysis [FE104]

Introduction

Econometric Principles and Data Analysis examines the interaction and confrontation between economic theory and economic data. It is concerned with the use of statistical and mathematical methods in analysing economic data, with the aim of providing economic theories with sufficient empirical foundation to enable them to be verified or refuted. Central attention is given to regression analysis – the major tool of statistical analysis in econometrics, to hypothesis testing and the treatment of heteroscedasticity and autocorrelation. The MICROFIT computer package is provided for regression analysis and diagnostic procedures.

Aims & Objectives

Econometrics is concerned with quantifying economic relations, with the provision of numerical estimates of the parameters involved and testing hypotheses embodied in economic relationships. This course aims to provide a basic introduction to econometric analysis, to enable students to examine economic theories with empirical data. In doing so, it examines the difficulties inherent in confronting theory with economic data in order to quantify economic relationships, in dealing with errors and problems in variables which can be only observed but not controlled, and the means of compensating for uncertainty in data.

Econometric Principles and Data Analysis is extended by the Part II option FE204 Econometric Analysis and Applications, which teaches more advanced techniques in quantitative methods. Econometric Principles and Data Analysis can be studied in its own right but normally the University would expect it to be taken as part of the MSc or Postgraduate Diploma in Financial Economics programme which provides the theoretical background required to interpret empirical data using statistical techniques.

Resources

Students receive a looseleaf binder containing eight ‘course units’; these texts are carefully structured to provide the main teaching and are equivalent to traditional course lectures, defining and exploring the main concepts and issues, locating these within current economics debate, introducing and linking the further assigned readings and setting out practical exercises for solution. Two obligatory assignments, which are marked by your CeFiMS tutors, and a specimen examination paper are also included within the student pack, along with the following:

Textbook:

Damodar N. Gujarati, Essentials of Econometrics, Second Edition, 1998 (International edition), McGraw-Hill Book Company., ISBN0071163069.

Computer Software:

MICROFIT, a user-friendly econometrics software programme, with an accompanying ‘MICROFIT Guide’ on how to use the software, which can be run on virtually any computer which is IBM (PC) compatible.

Course Timetable:

This shows the linkage between the various components of the course and indicates the schedule for reading the texts, submitting assignments, etc.

Course Content

Unit 1 Introduction

This first unit introduces some basic ideas on econometrics as the application of statistical and mathematical methods to the analysis of economic data, and the theoretical cornerstone of econometric theory and practice, regression analysis. Centrally concerned with the concept of regression, it also teaches the practical techniques and uses of the scatter plot as a practical tool of empirical analysis; how to enter data in MICROFIT, and to use the commands to obtain a summary of descriptive statistics of a variable, make a scatter plot and create logarithms of variables.

Unit 2 The Classical Linear Regression Model

Unit 2 introduces the basic, two-variable, linear regression model. The assumptions underlying the classical linear regression model and the techniques for estimating the model parameters from sample data are explained. It teaches the elementary use of the least squares method; classical linear regression; the Gauss–Markov theorem; standard errors and goodness of fit of a regression line, in addition to the following practical skills: how to formulate a simple regression model; to use MICROFIT to estimate the model from sample data; to interpret regression output including the signs of coefficients and their standard errors and the coefficient of determination, R squared, and the standard error of the regression.

Unit 3 Hypothesis Testing

The focus of this unit is on statistical reasoning with respect to interval estimation and hypothesis testing. The purpose of statistical inference is not only to draw conclusions from a sample of data about its population but also to incorporate an estimate of the reliability of the conclusions. In addition to its examination of the conceptual basis of statistical reasoning, the unit’s aim is to enable students to be versatile with constructing confidence intervals for regression coefficients; carrying out the t-test on regression coefficients; reading and interpreting the usual reported results of simple regression analysis.

Unit 4 The Multiple Regression Model – Estimation, Hypothesis Tests and Multicollinearity

Linear regression models with more than one explanatory variable are introduced in Unit 4, which teaches students how to formulate a multiple regression model; to use MICROFIT to estimate the model from sample data; to interpret the values for the partial regression coefficients, the standard error of estimate, the two versions of the coefficient of determination, R squared and R bar squared; to test hypotheses about individual partial regression coefficients using t statistics and to understand p-values; to test joint hypotheses with F tests; to analyse the linear relationships between pairs of variables in the regression model using correlation coefficients and partial correlation coefficients.

Unit 5 Heteroscedasticity

This unit explores techniques for making progress when one of the assumptions of the general linear regression model does not hold. The teaching specifically addresses heteroscedasticity — its causes, consequences and cures – by investigating four relevant questions: ‘what is heteroscedasticity?’, ‘does it matter?’, ‘how is it detected?’ and ‘what can be done about it?’. It summarises the central ideas of this important concept and how to overcome it, introducing the method of weighted least squares.

Unit 6 Autocorrelation

One of the assumptions of the classical regression model is that the disturbances in the model are not ‘autocorrelated’. Just as with heteroscedasticity, the presence of autocorrelation can have serious effects on the properties of estimators and test statistics. The nature of autorcorrelation is considered in this unit, along with tests and methods of dealing with such disturbances. The tests for autocorrelation considered include Residual Plots, the runs test, the Durbin-Watson test and the h test; the unit also demonstrates how to implement the Cochrane-Orcutt procedure in MICROFIT.

Unit 7 Nonnormal Disturbances

This unit focuses on another assumption of the classical normal linear regression model, the assumption of normally distributed disturbances. This has important consequences for hypothesis tests. The unit discusses the nature of normally distributed disturbances and use residual plots to compare cases of normal and nonnormal disturbances. the University examine the consequences of ignoring nonnormal disturbances, look at tests to detect nonnormality, and discuss what can be done about it.

Unit 8 Model Selection and Course Summary

Issues concerned with the specification of econometric models are covered in this final unit, which also reviews the course and introduces the more advanced econometrics techniques which are taught in the course Economic Analysis and Application. Unit 8 examines the characteristics of a ‘good’ model, one that explains the complex interactions within an economy, or within particular sectors of the economy, and in doing so, it highlights the problems of model building and how to avoid or correct them: the types of specification error which might be made; the effects of omitting a relevant variable and including an irrelevant variable; how to check for misspecification using residual plots, the Durbin-Watson statistic and Ramsey’s RESET test.

Tuition & Assessment

The course is assessed by two assignments and a three-hour examination held in the autumn. Each assignment consists of compulsory questions or problems to be solved, some of them using MICROFIT.