Universidade NOVA de Lisboa Faculdade de Economia 2009/2010 Principles of Econometrics Introduction 1
Textbook Wooldridge, Jeffrey M. Introductory Econometrics: A Modern Approach, 3 rd Edition, Thomson South-Western College Publishing, USA Further materials in www.docentes.fe.unl.pt/~azevedoj Homework Selected problems Computer exercises using Eviews Introduction 2
Grading Final grade: Analysis of a dataset (weighted 25%) Midterm (weighted 25%) Final exam (weighted 50%) Need a grade higher or equal to 8.0 (NOT 7.5 or 7.9) in the final exam. Exams: Late February (Midterm, in class) March 20 (Final) Introduction 3
Econometrics? The Econometric Society: An international society for the advancement of economic theory in its relation to statistics and mathematics Introduction 4
Why study Econometrics? Quantify economic relations, test and inform economic theories (e.g., neutrality of money) Policy evaluation (e.g., impact of a government retraining program on the wages of workers) Forecasting (GDP growth, unemployment, inflation) Need to use non-experimental, or observational data to make inferences Introduction 5
Types of Data Cross Sectional Each observation is an individual, firm, etc. at some point in time Cross-sectional data can be seen as a random sample If the data is not a random sample, we have a sample-selection problem Cross - Section Introduction 6
Types of Data Time Series and Panel Data Time series data is observed in various time periods e.g. stock prices, interest rates Not a random sample (in general), different problems to consider Time Series Can follow various random individuals over time panel data or longitudinal data (Cross Section Time Series) Introduction 7
A Simple Model: Returns to Education A model of human capital investment implies that more education should lead to higher earnings Consider the model: Earnings β 0 + β education + = 1 u Earnings is the dependent variable (or explained variable, or response variable, or the regressand) education is the independent variable (or explanatory variable, or control variable, or the regressor) u is the error term or disturbance, it should be unrelated to education. β 1 is the slope parameter and β 0 is the intercept parameter Introduction 8
Example: (continued) Earnings β 0 + β education + = 1 u The estimate of β 1 (based on the assumption that u is unrelated to education) is the return to education, but does it represent the ceteris paribus (everything else constant) effect of an additional year of education? The error term, u, includes other factors affecting earnings (e.g., natural ability that affects both education AND Earnings). Want to control for these factors as much as possible Some relevant variables are still unobserved (e.g., ability) Introduction 9
Example: Crime and Police CrimeRate β 0 + β Policeman + = 1 u The estimate ofβ 1 (based on the assumption that u is unrelated to Policeman) measures the effect of one additional policeman on the crime rate, but does it represent the ceteris paribus (everything else constant) effect of an additional policeman on the crime rate? The error term, u, includes other factors affecting crime (e.g., a new and popular illegal drug that can lead to both an increase in the police force, for political reasons, as well as an increase in crime rates). So, the number of policeman can vary due to variations in the crime rate itself! Causality is established only by theory, NOT by an Econometric exercise Introduction 10