Econometrics: Chapter 1: Introduction
Econometrics: Chapter 1: Introduction
Chapter 1: Introduction
Econometrics
Estimate the magnitude of quantitative relationships among economic variables Test economic hypotheses Forecast future outcomes
Uses of econometrics
political science, psychology, history, biology, medical research, and financial analysis.
are the predicted qualitative effects economically significant? Keynesian vs. monetarist vs. new classical models
Market research: finding the parameters of demand or supply curves to compute ownprice, cross-price, income, and supply elasticities
determining the return to a college degree How do alternative government policies affect economic outcomes?
How does labor supply, employment, and investment respond to changes in taxes and other policy variables How will a carbon tax affect pollution levels? How will changes in penalties or conviction rates affect crime rates? How will changing welfare rules affect income distribution, employment, poverty, and health care
Linear models
Intercept Slope
Scattergram
Scattergram = plot of observed data points
Error terms
observed data points do not fall exactly along a linear relationship an error term, ui, is introduced to account for these deviations from a linear relationship: Yi = bo+ b1Xi + ui
captures the effect of other variables not included in the model element of randomness in individual behavior incorrect measure of variables fitted equation may not reflect the actual functional relationship
Error terms
dependent variable is assumed to be determined by the independent variable(s) dependent variable is on the left-hand side of a regression equation independent variables appear on the right-hand side of a regression equation
Fitted values
Fitted value = value of the dependent variable predicted by the estimated regression line at a particular level of the independent variable Fitted values are denoted with a hat over the variable (e.g., )
Estimated parameters
Estimated parameters are also represented using a hat over the corresponding model parameter For example:
Intercept:
Slope:
= marginal propensity to consume
If rejected, formulate a new hypothesis If the hypothesis fails to be rejected, test the hypothesis again
Hypothesis testing
Hypothesis testing generally involves testing predictions concerning the magnitude or signs of model coefficients
Demand equation
Each slope coefficient represents the ceteris paribus effect of a change in that particular variable. Possible hypothesis tests?
Are price and quantity demanded inversely related? Are goods X and Y (or X and Z) substitutes and complements? Is this a normal or inferior good?
Forecasting
Demand forecasting Macroeconomic forecasting Predicting the outcome of elections Predicting outcomes of alternative policies
Model specification
endogenous variables
exogenous variables
Initially, it is assumed that the dependent variable is the only endogenous variable (all right-hand side variables assumed to be exogenous).
Specification error
MV = PY
Types of data
Cross-sectional
Time-series
Pooled
Assignment
Pp. 27-29: 3, 4, 7, 11